From d00784f5641a525297689adc337e16f6142412c5 Mon Sep 17 00:00:00 2001 From: neutrino2211 Date: Mon, 15 Dec 2025 21:47:46 +0000 Subject: [PATCH] Add malik/all_insights.json --- malik/all_insights.json | 15311 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 15311 insertions(+) create mode 100644 malik/all_insights.json diff --git a/malik/all_insights.json b/malik/all_insights.json new file mode 100644 index 0000000..1f378b5 --- /dev/null +++ b/malik/all_insights.json @@ -0,0 +1,15311 @@ +{ + "ABDULMALEEK-FINAL-YEAR-PROJECT": { + "author_metadata": "Abdul-Malik Abdullahi Mustapha. A final year Mechatronics Engineering student at the Federal University of Minna, Nigeria.", + "source_metadata": "Abdul-Malik final year project", + "insights": [ + { + "type": "fact", + "insight": "Document title: Artificial Intelligence-Enabled Multi-Function Activity Monitoring and Reporting System", + "content": "ARTIFICIAL INTELLIGENCE-ENABLED MULTI-FUNCTION ACTIVITY MONITORING AND REPORTING SYSTEM", + "attributes": [ + { + "attribute": "section", + "value": "title page" + }, + { + "attribute": "page_number", + "value": "1" + }, + { + "attribute": "document_id", + "value": "2016/1/60695ET" + }, + { + "attribute": "author_line", + "value": "MUSTAPHA, ABDUL-MALIK ABDULLAHI" + } + ] + }, + { + "type": "fact", + "insight": "Author line as shown: MUSTAPHA, ABDUL-MALIK ABDULLAHI", + "content": "MUSTAPHA, ABDUL-MALIK ABDULLAHI 2016/1/60695ET", + "attributes": [ + { + "attribute": "author_line", + "value": "MUSTAPHA, ABDUL-MALIK ABDULLAHI" + } + ] + }, + { + "type": "fact", + "insight": "Author full name (parsed): Abdul-Malik Abdullahi Mustapha", + "content": "MUSTAPHA, ABDUL-MALIK ABDULLAHI", + "attributes": [ + { + "attribute": "author_last_name", + "value": "Mustapha" + }, + { + "attribute": "author_given_names", + "value": "Abdul-Malik Abdullahi" + } + ] + }, + { + "type": "fact", + "insight": "Affiliation: Department of Mechatronics Engineering, School of Electrical Engineering and Technology, Federal University of Technology, Minna, Niger State", + "content": "DEPARTMENT OF MECHATRONICS ENGINEERING SCHOOL OF ELECTRICAL ENGINEERING AND TECHNOLOGY, FEDERAL UNIVERSITY OF TECHNOLOGY, MINNA, NIGER STATE", + "attributes": [ + { + "attribute": "department", + "value": "Mechatronics Engineering" + }, + { + "attribute": "school", + "value": "Electrical Engineering and Technology (School)" + }, + { + "attribute": "institution", + "value": "Federal University of Technology, Minna" + }, + { + "attribute": "location", + "value": "Minna, Niger State" + } + ] + }, + { + "type": "fact", + "insight": "Date on page: February 2022", + "content": "FEBRUARY 2022", + "attributes": [ + { + "attribute": "date_published", + "value": "February 2022" + } + ] + }, + { + "type": "fact", + "insight": "Page number on page: 1", + "content": "1", + "attributes": [ + { + "attribute": "page_number", + "value": "1" + } + ] + }, + { + "type": "comment", + "insight": "This page appears to be the title/cover page of an academic document (thesis/dissertation or project report) from the Federal University of Technology, Minna, Nigeria.", + "content": "ARTIFICIAL INTELLIGENCE-ENABLED MULTI-FUNCTION ACTIVITY MONITORING AND REPORTING SYSTEM MUSTAPHA, ABDUL-MALIK ABDULLAHI 2016/1/60695ET DEPARTMENT OF MECHATRONICS ENGINEERING SCHOOL OF ELECTRICAL ENGINEERING AND TECHNOLOGY, FEDERAL UNIVERSITY OF TECHNOLOGY, MINNA, NIGER STATE FEBRUARY 2022", + "attributes": [ + { + "attribute": "section", + "value": "title page" + }, + { + "attribute": "document_type_guess", + "value": "thesis/dissertation/project report title page" + } + ] + }, + { + "type": "fact", + "insight": "The project title is 'Artificial Intelligence-Enabled Multi-function Activity Monitoring and Reporting System'.", + "content": "CERTIFICATION This project, \"Artificial Intelligence-Enabled Multi-function Activity Monitoring and Reporting System\" by Mustapha Abdul-malik Abdullahi (2016/1/60695ET), satisfies the rules regulating the award of the Degree of Bachelor of Engineering (B.Eng.) at the Federal University of Technology, Minna, and it is authorized for the contribution of scientific knowledge and literary presentation.", + "attributes": [ + { + "attribute": "section", + "value": "Certification page" + }, + { + "attribute": "project_title", + "value": "Artificial Intelligence-Enabled Multi-function Activity Monitoring and Reporting System" + }, + { + "attribute": "source", + "value": "Page 2 of 56" + }, + { + "attribute": "author", + "value": "Mustapha Abdul-malik Abdullahi" + }, + { + "attribute": "author_id", + "value": "2016/1/60695ET" + }, + { + "attribute": "degree", + "value": "Bachelor of Engineering (B.Eng.)" + }, + { + "attribute": "university", + "value": "Federal University of Technology, Minna" + } + ] + }, + { + "type": "fact", + "insight": "The author is Mustapha Abdul-malik Abdullahi with the identifier 2016/1/60695ET.", + "content": "CERTIFICATION This project, \"Artificial Intelligence-Enabled Multi-function Activity Monitoring and Reporting System\" by Mustapha Abdul-malik Abdullahi (2016/1/60695ET), satisfies the rules regulating the award of the Degree of Bachelor of Engineering (B.Eng.) at the Federal University of Technology, Minna, and it is authorized for the contribution of scientific knowledge and literary presentation.", + "attributes": [ + { + "attribute": "section", + "value": "Certification page" + }, + { + "attribute": "author", + "value": "Mustapha Abdul-malik Abdullahi" + }, + { + "attribute": "author_id", + "value": "2016/1/60695ET" + }, + { + "attribute": "source", + "value": "Page 2 of 56" + }, + { + "attribute": "note", + "value": "Part of header content confirming authorship" + } + ] + }, + { + "type": "fact", + "insight": "The degree and institution are specified as Bachelor of Engineering (B.Eng.) at Federal University of Technology, Minna.", + "content": "CERTIFICATION This project, \"Artificial Intelligence-Enabled Multi-function Activity Monitoring and Reporting System\" by Mustapha Abdul-malik Abdullahi (2016/1/60695ET), satisfies the rules regulating the award of the Degree of Bachelor of Engineering (B.Eng.) at the Federal University of Technology, Minna, and it is authorized for the contribution of scientific knowledge and literary presentation.", + "attributes": [ + { + "attribute": "section", + "value": "Certification page" + }, + { + "attribute": "degree", + "value": "Bachelor of Engineering (B.Eng.)" + }, + { + "attribute": "university", + "value": "Federal University of Technology, Minna" + } + ] + }, + { + "type": "fact", + "insight": "The certification states that the project satisfies the rules regulating the award of the degree and is authorized for contribution of knowledge and literary presentation.", + "content": "CERTIFICATION This project, \"Artificial Intelligence-Enabled Multi-function Activity Monitoring and Reporting System\" by Mustapha Abdul-malik Abdullahi (2016/1/60695ET), satisfies the rules regulating the award of the Degree of Bachelor of Engineering (B.Eng.) at the Federal University of Technology, Minna, and it is authorized for the contribution of scientific knowledge and literary presentation.", + "attributes": [ + { + "attribute": "section", + "value": "Certification page" + }, + { + "attribute": "note", + "value": "Satisfies rules for degree award; authorized for knowledge and literary presentation" + } + ] + }, + { + "type": "fact", + "insight": "The page lists Engr. K.E. Jack as Project Supervisor and Prof. J.G. Kolo as Head of Department with corresponding signature lines.", + "content": "CERTIFICATION This project, \"Artificial Intelligence-Enabled Multi-function Activity Monitoring and Reporting System\" by Mustapha Abdul-malik Abdullahi (2016/1/60695ET), satisfies the rules regulating the award of the Degree of Bachelor of Engineering (B.Eng.) at the Federal University of Technology, Minna, and it is authorized for the contribution of scientific knowledge and literary presentation. Engr. K.E. Jack ……………………… Project Supervisor Signature and Date Prof. J.G. Kolo …………………... Head of Department. Signature and Date ", + "attributes": [ + { + "attribute": "section", + "value": "Certification page" + }, + { + "attribute": "supervisor", + "value": "Engr. K.E. Jack" + }, + { + "attribute": "hod", + "value": "Prof. J.G. Kolo" + }, + { + "attribute": "signature_lines", + "value": "Signature and Date" + } + ] + }, + { + "type": "fact", + "insight": "This is page 2 of 56 pages in the document (Certification page).", + "content": "CERTIFICATION This project, \"Artificial Intelligence-Enabled Multi-function Activity Monitoring and Reporting System\" by Mustapha Abdul-malik Abdullahi (2016/1/60695ET), satisfies the rules regulating the award of the Degree of Bachelor of Engineering (B.Eng.) at the Federal University of Technology, Minna, and it is authorized for the contribution of scientific knowledge and literary presentation. Engr. K.E. Jack ……………………… Project Supervisor Signature and Date Prof. J.G. Kolo …………………... Head of Department. Signature and Date ", + "attributes": [ + { + "attribute": "section", + "value": "Certification page" + }, + { + "attribute": "page_number", + "value": "2 of 56" + } + ] + }, + { + "type": "comment", + "insight": "This page contains signature placeholders (Signature and Date) for both the Project Supervisor and the Head of Department, indicating formal approval pending actual dates.", + "content": "CERTIFICATION This project, \"Artificial Intelligence-Enabled Multi-function Activity Monitoring and Reporting System\" by Mustapha Abdul-malik Abdullahi (2016/1/60695ET), satisfies the rules regulating the award of the Degree of Bachelor of Engineering (B.Eng.) at the Federal University of Technology, Minna, and it is authorized for the contribution of scientific knowledge and literary presentation. Engr. K.E. Jack ……………………… Project Supervisor Signature and Date Prof. J.G. Kolo …………………... Head of Department. Signature and Date ", + "attributes": [ + { + "attribute": "section", + "value": "Certification page" + }, + { + "attribute": "signature_lines", + "value": "Signature and Date" + } + ] + }, + { + "type": "comment", + "insight": "Dedication section states the project is dedicated to Allah and to people who helped during the university years.", + "content": "DEDICATION This project is dedicated to Allah for His faithfulness to me during my university years and to everyone who has helped in any manner over the years.", + "attributes": [ + { + "attribute": "section", + "value": "Dedication" + }, + { + "attribute": "page_number", + "value": "3" + }, + { + "attribute": "tone", + "value": "devotional/religious" + }, + { + "attribute": "topic", + "value": "Project dedication text" + }, + { + "attribute": "source_document", + "value": "Document page 3 of 56" + } + ] + }, + { + "type": "fact", + "insight": "Dedication to Allah and acknowledgement of those who helped during university years.", + "content": "DEDICATION This project is dedicated to Allah for His faithfulness to me during my university years and to everyone who has helped in any manner over the years.", + "attributes": [ + { + "attribute": "section", + "value": "DEDICATION" + }, + { + "attribute": "source_page", + "value": "4" + }, + { + "attribute": "author", + "value": "thesis author" + } + ] + }, + { + "type": "fact", + "insight": "Author spent more than five years at the Federal University of Technology in Minna.", + "content": "All thanks to Allah (S.W.T) for being Al-Rahman (The Merciful) and Al-Fattah (The Opener) during the more than five years I spent at the Federal University of Technology in Minna.", + "attributes": [ + { + "attribute": "section", + "value": "ACKNOWLEDGMENT" + }, + { + "attribute": "source_page", + "value": "4" + } + ] + }, + { + "type": "fact", + "insight": "Engr. Dr. K. E. Jack is acknowledged as supervisor and mentor.", + "content": "I wish to express my profound gratitude to Engr. Dr. K. E. Jack, my supervisor, for the mentorship he provided as both my instructor and my boss.", + "attributes": [ + { + "attribute": "person", + "value": "Engr. Dr. K. E. Jack" + }, + { + "attribute": "role", + "value": "supervisor/instructor/mentor" + }, + { + "attribute": "section", + "value": "ACKNOWLEDGMENT" + } + ] + }, + { + "type": "fact", + "insight": "Level advisers are Dr. Muhammad Enagi Bima, Dr. T. A. Folorunso, and Engr. Justice C. Annunso for their support and counsel.", + "content": "I'm appreciative My level advisers from 100 level to 500 level, Dr. Muhammad Enagi Bima, Dr. T. A. Folorunso, and Engr. Justice C. Annunso, true heroes as well for their support and counsel.", + "attributes": [ + { + "attribute": "section", + "value": "ACKNOWLEDGMENT" + }, + { + "attribute": "subjects", + "value": "level advisers (100 to 500 level)" + } + ] + }, + { + "type": "fact", + "insight": "Head of Department Prof. J.G. Kolo is thanked for tireless efforts to improve the department.", + "content": "I also want to thank the Head of the Department, Prof. J.G. Kolo, for his tireless efforts to make the department an outstanding one.", + "attributes": [ + { + "attribute": "person", + "value": "Prof. J.G. Kolo" + }, + { + "attribute": "role", + "value": "Head of Department" + }, + { + "attribute": "section", + "value": "ACKNOWLEDGMENT" + } + ] + }, + { + "type": "fact", + "insight": "Author expresses gratitude to their parents for support and enabling the journey.", + "content": "And to my amazing parents, I'd want to express my gratitude for always being there for me, for making this trip possible, and for their love, prayers, and support.", + "attributes": [ + { + "attribute": "relation", + "value": "parents" + }, + { + "attribute": "section", + "value": "ACKNOWLEDGMENT" + } + ] + }, + { + "type": "fact", + "insight": "Thanks extended to Sule, Sister, Musty, and the rest of the family and friends for support.", + "content": "I really thank you for this and everything else. Thank you to Sule, Sister, Musty, and the rest of my loving family and friends for being such a wonderful support system.", + "attributes": [ + { + "attribute": "subject", + "value": "Sule, Sister, Musty, family and friends" + }, + { + "attribute": "section", + "value": "ACKNOWLEDGMENT" + } + ] + }, + { + "type": "opinion", + "insight": "Comment acknowledging the importance of the Head of Department's role.", + "content": "Your role, whether direct or indirect, cannot be understated.", + "attributes": [ + { + "attribute": "section", + "value": "ACKNOWLEDGMENT" + }, + { + "attribute": "subject", + "value": "Head of Department" + } + ] + }, + { + "type": "comment", + "insight": "Document formatting/typographical notes observed in the page (line breaks, punctuation).", + "content": "The page shows spacing and punctuation inconsistencies as seen in the scanned text (e.g., stray line breaks and spacing).", + "attributes": [ + { + "attribute": "section", + "value": "FORMAT" + }, + { + "attribute": "source_page", + "value": "4" + } + ] + }, + { + "type": "comment", + "insight": "This page is the Acknowledgment section of a formal document, opening with religious thanks to Allah and gratitude to several individuals.", + "content": "ACKNOWLEDGMENT All thanks to Allah (S.W.T) for being Al-Rahman (The Merciful) and Al-Fattah (The Opener) during the more than five years I spent at the Federal University of Technology in Minna. All glory is due to Him.", + "attributes": [ + { + "attribute": "section", + "value": "Acknowledgment" + }, + { + "attribute": "page", + "value": "5" + }, + { + "attribute": "tone", + "value": "religious/reverential" + }, + { + "attribute": "source", + "value": "Document content" + } + ] + }, + { + "type": "fact", + "insight": "The author identifies Engr. Dr. K. E. Jack as supervisor and mentor.", + "content": "I wish to express my profound gratitude to Engr. Dr. K. E. Jack, my supervisor, for the mentorship he provided as both my instructor and my boss.", + "attributes": [ + { + "attribute": "person", + "value": "Engr. Dr. K. E. Jack" + }, + { + "attribute": "role", + "value": "supervisor, mentor, instructor, boss" + }, + { + "attribute": "section", + "value": "Acknowledgment" + } + ] + }, + { + "type": "fact", + "insight": "The author lists level advisers by name: Dr. Muhammad Enagi Bima, Dr. T. A. Folorunso, and Engr. Justice C. Annunso.", + "content": "I'm appreciative My level advisers from 100 level to 500 level, Dr. Muhammad Enagi Bima, Dr. T. A. Folorunso, and Engr. Justice C. Annunso, true heroes as well for their support and counsel.", + "attributes": [ + { + "attribute": "section", + "value": "Acknowledgment" + }, + { + "attribute": "people", + "value": "Dr. Muhammad Enagi Bima; Dr. T. A. Folorunso; Engr. Justice C. Annunso" + }, + { + "attribute": "roles", + "value": "level advisers" + }, + { + "attribute": "source", + "value": "Document content" + } + ] + }, + { + "type": "fact", + "insight": "The author thanks the Head of the Department, Prof. J.G. Kolo, for tireless efforts to make the department outstanding.", + "content": "I also want to thank the Head of the Department, Prof. J.G. Kolo, for his tireless efforts to make the department an outstanding one.", + "attributes": [ + { + "attribute": "section", + "value": "Acknowledgment" + }, + { + "attribute": "person", + "value": "Prof. J.G. Kolo" + }, + { + "attribute": "role", + "value": "Head of Department" + }, + { + "attribute": "tone", + "value": "positive/commendatory" + } + ] + }, + { + "type": "opinion", + "insight": "The phrase dismissing the opposite possibility, 'Your role, whether direct or indirect, cannot be understated,' conveys a strong positive assessment of the recipient's impact.", + "content": "Your role, whether direct or indirect, cannot be understated.", + "attributes": [ + { + "attribute": "section", + "value": "Acknowledgment" + }, + { + "attribute": "content", + "value": "Your role, whether direct or indirect, cannot be understated." + }, + { + "attribute": "tone", + "value": "positive/commendatory" + } + ] + }, + { + "type": "comment", + "insight": "The author expresses gratitude to their parents and mentions support from family; this reflects a personal and relational tone.", + "content": "And to my amazing parents, I'd want to express my gratitude for always being there for me, for making this trip possible, and for their love, prayers, and support.", + "attributes": [ + { + "attribute": "section", + "value": "Acknowledgment" + }, + { + "attribute": "relation", + "value": "parents" + } + ] + }, + { + "type": "comment", + "insight": "The author thanks Sule, Sister, Musty, and the rest of the loving family and friends for support.", + "content": "Thank you to Sule, Sister, Musty, and the rest of my loving family and friends for being such a wonderful support system.", + "attributes": [ + { + "attribute": "section", + "value": "Acknowledgment" + }, + { + "attribute": "relation", + "value": "family and friends" + } + ] + }, + { + "type": "fact", + "insight": "System architecture uses Raspberry Pi as microcontroller.", + "content": "ABSTRACT In the modern technological civilization, data is the new oil. The influence of \nefficient\n \n data\n has\n \naltered\n \nperformance\n \nstandards\n \nin\nterms\n \nof\n \naccuracy\n \nand\n \nspeed.\n \nEvery\n \nhospital\n \nand\n \nhousehold\n \nmust\n \nhave\n \na\n monitor­ing\n \nsystem\n \nas\n it\n \nis\n \nso\n \ncrucial.\n \nAs\n \na\n \nRaspberry\n \nPi\n \nis\n \nused\n \nin\n \nthe\n \nsystem's\n \narchitecture\n \nas\n \the\n \nmicrocontroller,\n \nwhile\n \ta\n \nmobile\n \napplication\n \nis\n \ndeveloped\n \nfor\n \nmonitoring\n \nactivities\n \ndetected\n \nin\n \nthe\n \nsystem\n \nto\n \nalert\n \nthe\n \nuser\n \ntogether\n \nw ith\n \na\n \nuser-controlled\n \nvideo\n \nrecording\n \nsystem.\n \nThe\n \nimprovement\n \nmay\n \nbe\n \nseen\n \nsince\n \ndata\n \nprocessing\n \nwas\n \nhandled\n \nby\n \ntwo\n \ncurrent\n \nindustry\n \nbuzzwords,\n \nComputer\n \nVision\n \n(CV)\n \nand\n \nArtificial\n \nIntelligence\n \n \n(AI).\n \nTwo\n \ntechnologies\n \nhave\n \nmade\n \nit\n \npossible\n \nto\n \ndo\n \nimportant\n \njobs\n \nlike\n \nobject\n \ndetecting\n \nsystems.\n \nAs\n \t the\n \nnumber\n \nof\n \nfeatures\n \nin\n \na\n \npicture\n \ngrows,\n \nso\n \ndoes\n \n \n\n the\n \nnecessity\n \nfor\n \nan\n \neffective\n \ntechnique\n \nto\n \nextract\n \nhidden\n \ninformation.\n \nThe\n \nCNN\n \nmodel\n \nis\n \nintended\n \nfor\n \nobject\n \ndetection\n \non\n \nthe\n \nCOCO\n \ndataset.\n \nAccuracy\n \nand\n \nprecision\n \nare\n \ntwo\n \nperformance\n \nmeasures\n \nthat\n \nare\n \nused\n \nto\n \nevaluate\n \nand\n \nverify\n \nthe\n \nmodel's\n \nperformance.\n \nSingle\n \nShot\n \nDetection\n \n(SSD),\n \nmobileNet,\n \nand\n \nPosenet\n \nare\n \nused\n \nto\n \nfollow\n \nobjects\n \nbetween\n \frames\n \nin\n \norder\n \nto\n \nestimate\n \ntheir\n \nposes\n \nvia\n \nthe\n \nvideo\n \nfeed.\n \nIt\n \nwas\n \nobserved\n \nthat\n \nthe\n \ndetection\n \nwas\n \neffective\n \nand\n \nefficient.\n \nFor\n \nreal-time\n \nactivity\n \ndetection\n \nand\n \nrecognition\n \napplications,\n \nthe\n \nalgorithms\n \nprovide\n \nreal-time,\n \naccurate,\n \nexact\n \nidentifications.\n \nAccording\n \ntest\n \nfindings,\n \nthe\n \nsystem\n \ncan\n \ncollect\n \nan\n \naverage\n \nof\n \n10.7\n \nframes\n \nper\n \nsecond\n \nand\n \nhas\n \na\n \ncamera\n \nmovement\n \nreaction\n \ntime\n \nof\n \nless\n \nthan\n \none\n \nsecond\n \nwhen\n \nthere\n \nare\n \nno\n \nmore\n \nthan\n \n2\n \nclients\n \nin\n \na\n \nnetwork.\n \nThe\n \nmonitoring\n \nsystem\n \nbecomes\n \na\n \nclever\n \nand\n \ndependable\n \nsystem\n \nafter\n \nsuccessful\n \ntesting\n \nof\n \nthe\n \nrecording\n \nand\n \nmotion\n \ndetecting\n \nfeatures.\n \nKeywords: Artificial Intelligence; Computer Vision; Convolution Neural \n \nNetwork;Single\n \nShot\n \nDetection;\n \nMobileNet;\n \nObject\n \ndetection;\n \nRasberry-Pi.\n \n \n \n6 ", + "attributes": [ + { + "attribute": "section", + "value": "Abstract" + }, + { + "attribute": "component", + "value": "Raspberry Pi" + }, + { + "attribute": "role", + "value": "microcontroller" + }, + { + "attribute": "source", + "value": "Page 6 (Abstract)" + } + ] + }, + { + "type": "fact", + "insight": "Mobile application is developed for monitoring and alerts, with a user-controlled video recording system.", + "content": "ABSTRACT In the modern technological civilization, data is the new oil. The influence of \nefficient\n \n data\n has\n \naltered\n \nperformance\n standards\n \nin\nterms\n \nof\n \naccuracy\n \nand\n \nspeed.\n \nEvery\n \nhospital\n \nand\n \nhousehold\n \nis\n \nhave\n \na\n \nmonitoring\n \nsystem\n \nas\n \tit\n \nis\n \nso\n \ncrucial.\n \nAs\n \na\n \nRaspberry\n \nPi\n \nis\n \nused\n \nin\n \nthe\n \nsystem's\n \narchitecture\n \nas\n \the\n \nmicrocontroller,\n \nwhile\n \ta\n \nmobile\n \napplication\n \nis\n \ndeveloped\n \nfor\n \nmonitoring\n \nactivities\n \ndetected\n \nin\n \nthe\n \nsystem\n \nto\n \nalert\n \tthe\n \nuser\n \ntogether\n \nwith\n \na\n \nuser-controlled\n \nvideo\n \nrecording\n \nsystem.\n \nThe\n \nimprovement\n \nmay\n \nbe\n \nseen\n \nsince\n \ndata\n \nprocessing\n \nwas\n \nhandled\n \nby\n \ntwo\n \ncurrent\n \nindustry\n \nbuzzwords,\n \nComputer\n \nVision\n \n(CV)\n \nand\n \nArtificial\n \nIntelligence\n \n(AI).\n \nTwo\n \ntechnologies\n \nhave\n \nmade\n \nit\n \npossible\n \nto\n \ndo\n \nimportant\n \njobs\n \nlike\n \nobject\n \ndetecting\n \nsystems.\n \nAs\n \nthe\n \nnumber\n \nof\n \nfeatures\n \nin\n \na\n \npicture\n \ngrows,\n \nso\n \ndoes\n \nthe\n \nnecessity\n \nfor\n \nan\n \neffective\n \ttechnique\n \nto\n \nextract\n \nhidden\n \ninformation.\n \nThe\n \nCNN\n \nmodel\n \nis\n \nintended\n \nfor\n \nobject\n \ndetection\n \non\n \nthe\n \nCOCO\n \ndataset.\n \nAccuracy\n \nand\n \nprecision\n \nare\n \ntwo\n \nperformance\n \nmeasures\n \nthat\n \nare\n \nused\n \nto\n \nevaluate\n \nand\n \nverify\n \nthe\n \nmodel's\n \nperformance.\n \nSingle\n \nShot\n \nDetection\n \n(SSD),\n \nmobileNet,\n \nand\n \nPosenet\n \nare\n \nused\n \nto\n \nfollow\n \nobjects\n \bin\n \nframes\n \nin\n \norder\n \nto\n \nestimate\n \ntheir\n \npos es\n \nvia\n \nthe\n \nvideo\n \nfeed.\n \nIt\n \nwas\n \nobserved\n \nthat\n \ndetection\n \nwas\n \neffective\n \nand\n \nefficient.\n \nFor\n \nreal-time\n \ndetection\n \nand\n \nrecognition\n \napplications,\n \nthe\n \nalgorithms\n \nprovide\n \nreal-time,\n \naccurate,\n \nexact\n \nidentifications.\n \nAccording\n \ntest\n \nfindings,\n \nthe\n \nsystem\n \ncan\n \ncollect\n \nan\n \naverage\n \nof\n \n10.7\n \nframes\n \nper\n \nsecond\n \nand\n \nhas\n \na\n \ncamera\n \nmovement\n \nreaction\n \ntime\n \nof\n \nless\n \nthan\n \none\n \nsecond\n \nwhen\n \nthere\n \nare\n \nno\n \nmore\n \nthan\n \n2\n \nclients\n \nin\n \na\n \nnetwork.\n \nThe\n \nmonitoring\n \nsystem\n \becomes\n \na\n \nclever\n \nand\n \ndependable\n \nsystem\n \nafter\n \nsuccessful\n \ntesting\n \nof\n \nthe\n \nrecording\n \nand\n \nmotion\n \ndetecting\n \nfeatures.\n \nKeywords: Artificial Intelligence; Computer Vision; Convolution Neural \n \nNetwork;Single\n \nShot\n \nDetection;\n \nMobileNet;\n \nObject\n \ndetection;\n \nRasberry-Pi.\n \n \n \n6 ", + "attributes": [ + { + "attribute": "section", + "value": "Abstract" + }, + { + "attribute": "feature", + "value": "mobile monitoring app with alerting" + }, + { + "attribute": "video_recording", + "value": "user-controlled" + } + ] + }, + { + "type": "opinion", + "insight": "Improvements are attributed to the use of Computer Vision (CV) and Artificial Intelligence (AI).", + "content": "ABSTRACT In the modern technological civilization, data is the new oil. The influence of \nefficient\n \n data\n has\n \naltered\n \nperformance\n standards\n \nin\nterms\n \nof\n \naccuracy\n \nand\n \nspeed.\n \nEvery\n \nhospital\n \nand\n \nhousehold\n \nis\n \nhave\n \na\n \nmonitoring\n \nsystem\n \nas\n \tit\n \nis\n \nso\n \ncrucial.\n \nAs\n \na\n \nRaspberry\n \nPi\n \nis\n \nused\n \nin\n \nthe\n \nsystem's\n \narchitecture\n \nas\n \the\n \nmicrocontroller,\n \nwhile\n \ta\n \nmobile\n \napplication\n \nis\n \ndeveloped\n \nfor\n \nmonitoring\n \nactivities\n \ndetected\n \nin\n \nthe\n \nsystem\n \nto\n \nalert\n \tthe\n \nuser\n \ntogether\n \nwith\n \na\n \nuser-controlled\n \nvideo\n \nrecording\n \nsystem.\n \nThe\n \nimprovement\n \nmay\n \nbe\n \nseen\n \nsince\n \ndata\n \nprocessing\n \nwas\n \nhandled\n \nby\n \ntwo\n \ncurrent\n \nindustry\n \nbuzzwords,\n \nComputer\n \nVision\n \n(CV)\n \nand\n \nArtificial\n \nIntelligence\n \n(AI).\n \nTwo\n \ntechnologies\n \nhave\n \nmade\n \nit\n \npossible\n \nto\n \ndo\n \nimportant\n \njobs\n \nlike\n \nobject\n \ndetecting\n \nsystems.\n \nAs\n \nthe\n \nnumber\n \nof\n \nfeatures\n \nin\n \na\n \npicture\n \ngrows,\n \nso\n \ndoes\n \nthe\n \nnecessity\n \nfor\n \nan\n \neffective\n \ttechnique\n \nto\n \nextract\n \nhidden\n \ninformation.\n \nThe\n \nCNN\n \nmodel\n \nis\n \nintended\n \nfor\n \nobject\n \ndetection\n \non\n \nthe\n \nCOCO\n \ndataset.\n \nAccuracy\n \nand\n \nprecision\n \nare\n \ntwo\n \nperformance\n \nmeasures\n \nthat\n \nare\n \nused\n \nto\n \nevaluate\n \nand\n \nverify\n \nthe\n \nmodel's\n \nperformance.\n \nSingle\n \nShot\n \nDetection\n \n(SSD),\n \nmobileNet,\n \nand\n \nPosenet\n \nare\n \nused\n \nto\n \nfollow\n \nobjects\n \bin\n \nframes\n \nin\n \norder\n \nto\n \nestimate\n \ntheir\n \npos es\n \nvia\n \nthe\n \nvideo\n \nfeed.\n \nIt\n \nwas\n \nobserved\n \nthat\n \ndetection\n \nwas\n \neffective\n \nand\n \nefficient.\n \nFor\n \nreal-time\n \ndetection\n \nand\n \nrecognition\n \napplications,\n \tthe\n \nalgorithms\n \nprovide\n \nreal-time,\n \naccurate,\n \nexact\n \nidentifications.\n \nAccording\n \ntest\n \nfindings,\n \nthe\n \nsystem\n \ncan\n \ncollect\n \nan\n \naverage\n \nof\n \n10.7\n \nframes\n \nper\n \nsecond\n \nand\n \nhas\n \na\n \ncamera\n \nmovement\n \nreaction\n \ntime\n \nof\n \nless\n \nthan\n \none\n \nsecond\n \nwhen\n \nthere\n \nare\n \nno\n \nmore\n \nthan\n \n2\n \nclients\n \nin\n \na\n \nnetwork.\n \nThe\n \nmonitoring\n \nsystem\n \becomes\n \na\n \nclever\n \nand\n \ndependable\n \nsystem\n \nafter\n \nsuccessful\n \ntesting\n \nof\n \nthe\n \nrecording\n \nand\n \nmotion\n \ndetecting\n \nfeatures.\n \nKeywords: Artificial Intelligence; Computer Vision; Convolution Neural \n \nNetwork;Single\n \nShot\n \nDetection;\n \nMobileNet;\n \nObject\n \ndetection;\n \nRasberry-Pi.\n \n \n \n6 ", + "attributes": [ + { + "attribute": "section", + "value": "Abstract" + }, + { + "attribute": "topic", + "value": "CV/AI influence on improvements" + }, + { + "attribute": "tone", + "value": "positive" + } + ] + }, + { + "type": "fact", + "insight": "CNN model is intended for object detection on the COCO dataset.", + "content": "ABSTRACT In the modern technological civilization, data is the new oil. The influence of \nefficient\n \n data\n has\n \naltered\n \nperformance\n standards\n \nin\nterms\n \nof\n \naccuracy\n \nand\n \nspeed.\n \nEvery\n \nhospital\n \nand\n \nhousehold\n \nis\n \nhave\n \na\n \nmonitoring\n \nsystem\n \nas\n \tit\n \nis\n \nso\n \ncrucial.\n \nAs\n \na\n \nRaspberry\n \nPi\n \nis\n \nused\n \nin\n \nthe\n \nsystem's\n \narchitecture\n \nas\n \the\n \nmicrocontroller,\n \nwhile\n \ta\n \nmobile\n \napplication\n \nis\n \ndeveloped\n \nfor\n \nmonitoring\n \nactivities\n \ndetected\n \nin\n \nthe\n \nsystem\n \nto\n \nalert\n \tthe\n \nuser\n \ntogether\n \nwith\n \na\n \nuser-controlled\n \nvideo\n \nrecording\n \nsystem.\n \nThe\n \nimprovement\n \nmay\n \nbe\n \nseen\n \nsince\n \ndata\n \nprocessing\n \nwas\n \nhandled\n \nby\n \ntwo\n \ncurrent\n \nindustry\n \nbuzzwords,\n \nComputer\n \nVision\n \n(CV)\n \nand\n \nArtificial\n \nIntelligence\n \n(AI).\n \nTwo\n \ntechnologies\n \nhave\n \nmade\n \nit\n \npossible\n \nto\n \ndo\n \nimportant\n \njobs\n \nlike\n \nobject\n \ndetecting\n \nsystems.\n \nAs\n \nthe\n \nnumber\n \nof\n \nfeatures\n \nin\n \na\n \npicture\n \ngrows,\n \nso\n \ndoes\n \nthe\n \nnecessity\n \nfor\n \nan\n \neffective\n \ttechnique\n \nto\n \nextract\n \nhidden\n \ninformation.\n \nThe\n \nCNN\n \nmodel\n \nis\n \nintended\n \nfor\n \nobject\n \ndetection\n \non\n \nthe\n \nCOCO\n \ndataset.\n \nAccuracy\n \nand\n \nprecision\n \nare\n \ntwo\n \nperformance\n \nmeasures\n \nthat\n \nare\n \nused\n \nto\n \nevaluate\n \nand\n \nverify\n \nthe\n \nmodel's\n \nperformance.\n \nSingle\n \nShot\n \nDetection\n \n(SSD),\n \nmobileNet,\n \nand\n \nPosenet\n \nare\n \nused\n \nto\n \nfollow\n \nobjects\n \bin\n \nframes\n \nin\n \norder\n \nto\n \nestimate\n \ntheir\n \npos es\n \nvia\n \nthe\n \nvideo\n \nfeed.\n \nIt\n \nwas\n \nobserved\n \nthat\n \ndetection\n \nwas\n \neffective\n \nand\n \nefficient.\n \nFor\n \nreal-time\n \ndetection\n \nand\n \nrecognition\n \napplications,\n \tthe\n \nalgorithms\n \nprovide\n \nreal-time,\n \naccurate,\n \nexact\n \nidentifications.\n \nAccording\n \ntest\n \nfindings,\n \nthe\n \nsystem\n \ncan\n \ncollect\n \nan\n \naverage\n \nof\n \n10.7\n \nframes\n \nper\n \nsecond\n \nand\n \nhas\n \na\n \ncamera\n \nmovement\n \nreaction\n \ntime\n \nof\n \nless\n \nthan\n \none\n \nsecond\n \nwhen\n \nthere\n \nare\n \nno\n \nmore\n \nthan\n \n2\n \nclients\n \nin\n \na\n \nnetwork.\n \nThe\n \nmonitoring\n \nsystem\n \becomes\n \na\n \nclever\n \nand\n \ndependable\n \nsystem\n \nafter\n \nsuccessful\n \ntesting\n \nof\n \nthe\n \nrecording\n \nand\n \nmotion\n \ndetecting\n \nfeatures.\n \nKeywords: Artificial Intelligence; Computer Vision; Convolution Neural \n \nNetwork;Single\n \nShot\n \nDetection;\n \nMobileNet;\n \nObject\n \ndetection;\n \nRasberry-Pi.\n \n \n \n6 ", + "attributes": [ + { + "attribute": "section", + "value": "Abstract" + }, + { + "attribute": "model", + "value": "CNN for object detection" + }, + { + "attribute": "dataset", + "value": "COCO" + } + ] + }, + { + "type": "fact", + "insight": "The CNN model is intended for object detection on the COCO dataset.", + "content": "ABSTRACT In the modern technological civilization, data is the new oil. The influence of \nefficient\n \n data\n has\n \naltered\n \nperformance\n standards\n \nin\nterms\n \nof\n \naccuracy\n \nand\n \nspeed.\n \nEvery\n \nhospital\n \nand\n \nhousehold\n \nis\n \nhave\n \na\n \nmonitoring\n \nsystem\n \nas\n \tit\n \nis\n \nso\n \ncrucial.\n \nAs\n \na\n \nRaspberry\n \nPi\n \nis\n \nused\n \nin\n \nthe\n \nsystem's\n \narchitecture\n \nas\n \the\n \nmicrocontroller,\n \nwhile\n \ta\n \nmobile\n \napplication\n \nis\n \ndeveloped\n \nfor\n \nmonitoring\n \nactivities\n \ndetected\n \nin\n \nthe\n \nsystem\n \nto\n \nalert\n \tthe\n \nuser\n \ntogether\n \nwith\n \na\n \nuser-controlled\n \nvideo\n \nrecording\n \nsystem.\n \nThe\n \nimprovement\n \nmay\n \nbe\n \nseen\n \nsince\n \ndata\n \nprocessing\n \nwas\n \nhandled\n \nby\n \ntwo\n \ncurrent\n \nindustry\n \nbuzzwords,\n \nComputer\n \nVision\n \n(CV)\n \nand\n \nArtificial\n \nIntelligence\n \n(AI).\n \nTwo\n \ntechnologies\n \nhave\n \nmade\n \nit\n \npossible\n \nto\n \ndo\n \nimportant\n \njobs\n \nlike\n \nobject\n \ndetecting\n \nsystems.\n \nAs\n \nthe\n \nnumber\n \nof\n \nfeatures\n \nin\n \na\n \npicture\n \ngrows,\n \nso\n \ndoes\n \nthe\n \nnecessity\n \nfor\n \nan\n \neffective\n \ttechnique\n \nto\n \nextract\n \nhidden\n \ninformation.\n \nThe\n \nCNN\n \nmodel\n \nis\n \nintended\n \nfor\n \nobject\n \ndetection\n \non\n \nthe\n \nCOCO\n \ndataset.\n \nAccuracy\n \nand\n \nprecision\n \nare\n \ntwo\n \nperformance\n \nmeasures\n \nthat\n \nare\n \nused\n \nto\n \nevaluate\n \nand\n \nverify\n \nthe\n \nmodel's\n \nperformance.\n \nSingle\n \nShot\n \nDetection\n \n(SSD),\n \nmobileNet,\n \nand\n \nPosenet\n \nare\n \nused\n \nto\n \nfollow\n \nobjects\n \bin\n \nframes\n \nin\n \norder\n \nto\n \nestimate\n \ntheir\n \npos es\n \nvia\n \nthe\n \nvideo\n \nfeed.\n \nIt\n \nwas\n \nobserved\n \nthat\n \ndetection\n \nwas\n \neffective\n \nand\n \nefficient.\n \nFor\n \nreal-time\n \ndetection\n \nand\n \nrecognition\n \napplications,\n \tthe\n \nalgorithms\n \nprovide\n \nreal-time,\n \naccurate,\n \nexact\n \nidentifications.\n \nAccording\n \ntest\n \nfindings,\n \nthe\n \nsystem\n \ncan\n \ncollect\n \nan\n \naverage\n \nof\n \n10.7\n \nframes\n \nper\n \nsecond\n \nand\n \nhas\n \na\n \ncamera\n \nmovement\n \nreaction\n \ntime\n \nof\n \nless\n \nthan\n \none\n \nsecond\n \nwhen\n \nthere\n \nare\n \nno\n \nmore\n \nthan\n \n2\n \nclients\n \nin\n \na\n \nnetwork.\n \nThe\n \nmonitoring\n \nsystem\n \becomes\n \na\n \nclever\n \nand\n \ndependable\n \nsystem\n \nafter\n \nsuccessful\n \ntesting\n \nof\n \nthe\n \nrecording\n \nand\n \nmotion\n \ndetecting\n \nfeatures.\n \nKeywords: Artificial Intelligence; Computer Vision; Convolution Neural \n \nNetwork;Single\n \nShot\n \nDetection;\n \nMobileNet;\n \nObject\n \ndetection;\n \nRasberry-Pi.\n \n \n \n6 ", + "attributes": [ + { + "attribute": "section", + "value": "Abstract" + }, + { + "attribute": "model", + "value": "CNN for object detection" + }, + { + "attribute": "dataset", + "value": "COCO" + } + ] + }, + { + "type": "fact", + "insight": "The detection was observed to be effective and efficient.", + "content": "ABSTRACT In the modern technological civilization, data is the new oil. The influence of \nefficient\n \n data\n has\n \naltered\n \nperformance\n standards\n \nin\nterms\n \nof\n \naccuracy\n \nand\n \nspeed.\n \nEvery\n \nhospital\n \nand\n \nhousehold\n \nis\n \nhave\n \na\n \nmonitoring\n \nsystem\n \nas\n \tit\n \nis\n \nso\n \ncrucial.\n \nAs\n \na\n \nRaspberry\n \nPi\n \nis\n \nused\n \nin\n \nthe\n \nsystem's\n \narchitecture\n \nas\n \the\n \nmicrocontroller,\n \nwhile\n \ta\n \nmobile\n \napplication\n \nis\n \ndeveloped\n \nfor\n \nmonitoring\n \nactivities\n \ndetected\n \nin\n \nthe\n \nsystem\n \nto\n \nalert\n \tthe\n \nuser\n \ntogether\n \nwith\n \na\n \nuser-controlled\n \nvideo\n \nrecording\n \nsystem.\n \nThe\n \nimprovement\n \nmay\n \nbe\n \nseen\n \nsince\n \ndata\n \nprocessing\n \nwas\n \nhandled\n \nby\n \ntwo\n \ncurrent\n \nindustry\n \nbuzzwords,\n \nComputer\n \nVision\n \n(CV)\n \nand\n \nArtificial\n \nIntelligence\n \n(AI).\n \nTwo\n \ntechnologies\n \nhave\n \nmade\n \nit\n \npossible\n \nto\n \ndo\n \nimportant\n \njobs\n \nlike\n \nobject\n \ndetecting\n \nsystems.\n \nAs\n \nthe\n \nnumber\n \nof\n \nfeatures\n \nin\n \na\n \npicture\n \ngrows,\n \nso\n \ndoes\n \nthe\n \nnecessity\n \nfor\n \nan\n \neffective\n \ttechnique\n \nto\n \nextract\n \nhidden\n \ninformation.\n \nThe\n \nCNN\n \nmodel\n \nis\n \nintended\n \nfor\n \nobject\n \ndetection\n \non\n \nthe\n \nCOCO\n \ndataset.\n \nAccuracy\n \nand\n \nprecision\n \nare\n \ntwo\n \nperformance\n \nmeasures\n \nthat\n \nare\n \nused\n \nto\n \nevaluate\n \nand\n \nverify\n \nthe\n \nmodel's\n \nperformance.\n \nSingle\n \nShot\n \nDetection\n \n(SSD),\n \nmobileNet,\n \nand\n \nPosenet\n \nare\n \nused\n \nto\n \nfollow\n \nobjects\n \bin\n \nframes\n \nin\n \norder\n \nto\n \nestimate\n \ntheir\n \npos es\n \nvia\n \nthe\n \nvideo\n \nfeed.\n \nIt\n \nwas\n \nobserved\n \nthat\n \ndetection\n \nwas\n \neffective\n \nand\n \nefficient.\n \nFor\n \nreal-time\n \ndetection\n \nand\n \nrecognition\n \napplications,\n \tthe\n \nalgorithms\n \nprovide\n \nreal-time,\n \naccurate,\n \nexact\n \nidentifications.\n \nAccording\n \ntest\n \nfindings,\n \nthe\n \nsystem\n \ncan\n \ncollect\n \nan\n \naverage\n \nof\n \n10.7\n \nframes\n \nper\n \nsecond\n \nand\n \nhas\n \na\n \ncamera\n \nmovement\n \nreaction\n \ntime\n \nof\n \nless\n \nthan\n \none\n \nsecond\n \nwhen\n \nthere\n \nare\n \nno\n \nmore\n \nthan\n \n2\n \nclients\n \nin\n \na\n \nnetwork.\n \nThe\n \nmonitoring\n \nsystem\n \becomes\n \na\n \nclever\n \nand\n \ndependable\n \nsystem\n \nafter\n \nsuccessful\n \ntesting\n \nof\n \nthe\n \nrecording\n \nand\n \nmotion\n \ndetecting\n \nfeatures.\n \nKeywords: Artificial Intelligence; Computer Vision; Convolution Neural \n \nNetwork;Single\n \nShot\n \nDetection;\n \nMobileNet;\n \nObject\n \ndetection;\n \nRasberry-Pi.\n \n \n \n6 ", + "attributes": [ + { + "attribute": "section", + "value": "Abstract" + }, + { + "attribute": "topic", + "value": "Observation of detection effectiveness" + } + ] + }, + { + "type": "fact", + "insight": "For real-time activity detection and recognition, the algorithms provide real-time, accurate identifications.", + "content": "ABSTRACT In the modern technological civilization, data is the new oil. The influence of \nefficient\n \n data\n has\n \naltered\n \nperformance\n standards\n \nin\nterms\n \nof\n \naccuracy\n \nand\n \nspeed.\n \nEvery\n \nhospital\n \nand\n \nhousehold\n \nis\n \nhave\n \na\n \nmonitoring\n \nsystem\n \nas\n \tit\n \nis\n \nso\n \ncrucial.\n \nAs\n \na\n \nRaspberry\n \nPi\n \nis\n \nused\n \nin\n \nthe\n \nsystem's\n \narchitecture\n \nas\n \the\n \nmicrocontroller,\n \nwhile\n \ta\n \nmobile\n \napplication\n \nis\n \ndeveloped\n \nfor\n \nmonitoring\n \nactivities\n \ndetected\n \nin\n \nthe\n \nsystem\n \nto\n \nalert\n \tthe\n \nuser\n \ntogether\n \nwith\n \na\n \nuser-controlled\n \nvideo\n \nrecording\n \nsystem.\n \nThe\n \nimprovement\n \nmay\n \nbe\n \nseen\n \nsince\n \ndata\n \nprocessing\n \nwas\n \nhandled\n \nby\n \ntwo\n \ncurrent\n \nindustry\n \nbuzzwords,\n \nComputer\n \nVision\n \n(CV)\n \nand\n \nArtificial\n \nIntelligence\n \n(AI).\n \nTwo\n \ntechnologies\n \nhave\n \nmade\n \nit\n \npossible\n \nto\n \ndo\n \nimportant\n \njobs\n \nlike\n \nobject\n \ndetecting\n \nsystems.\n \nAs\n \nthe\n \nnumber\n \nof\n \nfeatures\n \nin\n \na\n \npicture\n \ngrows,\n \nso\n \ndoes\n \nthe\n \nnecessity\n \nfor\n \nan\n \neffective\n \ttechnique\n \nto\n \nextract\n \nhidden\n \ninformation.\n \nThe\n \nCNN\n \nmodel\n \nis\n \nintended\n \nfor\n \nobject\n \ndetection\n \non\n \nthe\n \nCOCO\n \ndataset.\n \nAccuracy\n \nand\n \nprecision\n \nare\n \ntwo\n \nperformance\n \nmeasures\n \nthat\n \nare\n \nused\n \nto\n \nevaluate\n \nand\n \nverify\n \nthe\n \nmodel's\n \nperformance.\n \nSingle\n \nShot\n \nDetection\n \n(SSD),\n \nmobileNet,\n \nand\n \nPosenet\n \nare\n \nused\n \nto\n \nfollow\n \nobjects\n \bin\n \nframes\n \nin\n \norder\n \nto\n \nestimate\n \ntheir\n \npos es\n \nvia\n \nthe\n \nvideo\n \nfeed.\n \nIt\n \nwas\n \nobserved\n \nthat\n \ndetection\n \nwas\n \neffective\n \nand\n \nefficient.\n \nFor\n \nreal-time\n \ndetection\n \nand\n \nrecognition\n \napplications,\n \tthe\n \nalgorithms\n \nprovide\n \nreal-time,\n \naccurate,\n \nexact\n \nidentifications.\n \nAccording\n \ntest\n \nfindings,\n \nthe\n \nsystem\n \ncan\n \ncollect\n \nan\n \naverage\n \nof\n \n10.7\n \nframes\n \nper\n \nsecond\n \nand\n \nhas\n \na\n \ncamera\n \nmovement\n \nreaction\n \ntime\n \nof\n \nless\n \nthan\n \none\n \nsecond\n \nwhen\n \nthere\n \nare\n \nno\n \nmore\n \nthan\n \n2\n \nclients\n \nin\n \na\n \nnetwork.\n \nThe\n \nmonitoring\n \nsystem\n \becomes\n \na\n \nclever\n \nand\n \ndependable\n \nsystem\n \nafter\n \nsuccessful\n \ntesting\n \nof\n \nthe\n \nrecording\n \nand\n \nmotion\n \ndetecting\n \nfeatures.\n \nKeywords: Artificial Intelligence; Computer Vision; Convolution Neural \n \nNetwork;Single\n \nShot\n \nDetection;\n \nMobileNet;\n \nObject\n \ndetection;\n \nRasberry-Pi.\n \n \n \n6 ", + "attributes": [ + { + "attribute": "section", + "value": "Abstract" + }, + { + "attribute": "summary", + "value": "real-time identifications" + } + ] + }, + { + "type": "opinion", + "insight": "The monitoring system becomes clever and dependable after successful testing of the recording and motion-detecting features.", + "content": "ABSTRACT In the modern technological civilization, data is the new oil. The influence of \nefficient\n \n data\n has\n \naltered\n \nperformance\n standards\n \nin\nterms\n \nof\n \naccuracy\n \nand\n \nspeed.\n \nEvery\n \nhospital\n \nand\n \nhousehold\n \nis\n \nhave\n \na\n \nmonitoring\n \nsystem\n \nas\n \tit\n \nis\n \nso\n \ncrucial.\n \nAs\n \na\n \nRaspberry\n \nPi\n \nis\n \nused\n \nin\n \nthe\n \nsystem's\n \narchitecture\n \nas\n \the\n \nmicrocontroller,\n \nwhile\n \ta\n \nmobile\n \napplication\n \nis\n \ndeveloped\n \nfor\n \nmonitoring\n \nactivities\n \ndetected\n \nin\n \nthe\n \nsystem\n \nto\n \nalert\n \tthe\n \nuser\n \ntogether\n \nwith\n \na\n \nuser-controlled\n \nvideo\n \nrecording\n \nsystem.\n \nThe\n \nimprovement\n \nmay\n \nbe\n \nseen\n \nsince\n \ndata\n \nprocessing\n \nwas\n \nhandled\n \nby\n \ntwo\n \ncurrent\n \nindustry\n \nbuzzwords,\n \nComputer\n \nVision\n \n(CV)\n \nand\n \nArtificial\n \nIntelligence\n \n(AI).\n \nTwo\n \ntechnologies\n \nhave\n \nmade\n \nit\n \npossible\n \nto\n \ndo\n \nimportant\n \njobs\n \nlike\n \nobject\n \ndetecting\n \nsystems.\n \nAs\n \nthe\n \nnumber\n \nof\n \nfeatures\n \nin\n \na\n \npicture\n \ngrows,\n \nso\n \ndoes\n \nthe\n \nnecessity\n \nfor\n \nan\n \neffective\n \ttechnique\n \nto\n \nextract\n \nhidden\n \ninformation.\n \nThe\n \nCNN\n \nmodel\n \nis\n \nintended\n \nfor\n \nobject\n \ndetection\n \non\n \nthe\n \nCOCO\n \ndataset.\n \nAccuracy\n \nand\n \nprecision\n \nare\n \ntwo\n \nperformance\n \nmeasures\n \nthat\n \nare\n \nused\n \nto\n \nevaluate\n \nand\n \nverify\n \nthe\n \nmodel's\n \nperformance.\n \nSingle\n \nShot\n \nDetection\n \n(SSD),\n \nmobileNet,\n \nand\n \nPosenet\n \nare\n \nused\n \nto\n \nfollow\n \nobjects\n \bin\n \nframes\n \nin\n \norder\n \nto\n \nestimate\n \ntheir\n \npos es\n \nvia\n \nthe\n \nvideo\n \nfeed.\n \nIt\n \nwas\n \nobserved\n \nthat\n \ndetection\n \nwas\n \neffective\n \nand\n \nefficient.\n \nFor\n \nreal-time\n \ndetection\n \nand\n \nrecognition\n \napplications,\n \tthe\n \nalgorithms\n \nprovide\n \nreal-time,\n \naccurate,\n \nexact\n \nidentifications.\n \nAccording\n \ntest\n \nfindings,\n \nthe\n \nsystem\n \ncan\n \ncollect\n \nan\n \naverage\n \nof\n \n10.7\n \nframes\n \nper\n \nsecond\n \nand\n \nhas\n \na\n \ncamera\n \nmovement\n \nreaction\n \ntime\n \nof\n \nless\n \nthan\n \none\n \nsecond\n \nwhen\n \nthere\n \nare\n \nno\n \nmore\n \nthan\n \n2\n \nclients\n \nin\n \na\n \nnetwork.\n \nThe\n \nmonitoring\n \nsystem\n \becomes\n \na\n \nclever\n \nand\n \ndependable\n \nsystem\n \nafter\n \nsuccessful\n \ntesting\n \nof\n \nthe\n \nrecording\n \nand\n \nmotion\n \ndetecting\n \nfeatures.\n \nKeywords: Artificial Intelligence; Computer Vision; Convolution Neural \n \nNetwork;Single\n \nShot\n \nDetection;\n \nMobileNet;\n \nObject\n \ndetection;\n \nRasberry-Pi.\n \n \n \n6 ", + "attributes": [ + { + "attribute": "section", + "value": "Abstract" + }, + { + "attribute": "summary", + "value": "clever and dependable after testing" + } + ] + }, + { + "type": "opinion", + "insight": "The detection claims (real-time, accurate identifications) are presented as results of algorithms.", + "content": "ABSTRACT In the modern technological civilization, data is the new oil. The influence of \nefficient\n \n data\n has\n \naltered\n \nperformance\n standards\n \nin\nterms\n \nof\n \naccuracy\n \nand\n \nspeed.\n \nEvery\n \nhospital\n \nand\n \nhousehold\n \nis\n \nhave\n \na\n \nmonitoring\n \nsystem\n \nas\n \tit\n \nis\n \nso\n \ncrucial.\n \nAs\n \na\n \nRaspberry\n \nPi\n \nis\n \nused\n \nin\n \nthe\n \nsystem's\n \narchitecture\n \nas\n \the\n \nmicrocontroller,\n \nwhile\n \ta\n \nmobile\n \napplication\n \nis\n \ndeveloped\n \nfor\n \nmonitoring\n \nactivities\n \ndetected\n \nin\n \nthe\n \nsystem\n \nto\n \nalert\n \tthe\n \nuser\n \ntogether\n \nwith\n \na\n \nuser-controlled\n \nvideo\n \nrecording\n \nsystem.\n \nThe\n \nimprovement\n \nmay\n \nbe\n \nseen\n \nsince\n \ndata\n \nprocessing\n \nwas\n \nhandled\n \nby\n \ntwo\n \ncurrent\n \nindustry\n \nbuzzwords,\n \nComputer\n \nVision\n \n(CV)\n \nand\n \nArtificial\n \nIntelligence\n \n(AI).\n \nTwo\n \ntechnologies\n \nhave\n \nmade\n \nit\n \npossible\n \nto\n \ndo\n \nimportant\n \njobs\n \nlike\n \nobject\n \ndetecting\n \nsystems.\n \nAs\n \nthe\n \nnumber\n \nof\n \nfeatures\n \nin\n \na\n \npicture\n \ngrows,\n \nso\n \ndoes\n \nthe\n \nnecessity\n \nfor\n \nan\n \neffective\n \ttechnique\n \nto\n \nextract\n \nhidden\n \ninformation.\n \nThe\n \nCNN\n \nmodel\n \nis\n \nintended\n \nfor\n \nobject\n \ndetection\n \non\n \nthe\n \nCOCO\n \ndataset.\n \nAccuracy\n \nand\n \nprecision\n \nare\n \ntwo\n \nperformance\n \nmeasures\n \nthat\n \nare\n \nused\n \nto\n \nevaluate\n \nand\n \nverify\n \nthe\n \nmodel's\n \nperformance.\n \nSingle\n \nShot\n \nDetection\n \n(SSD),\n \nmobileNet,\n \nand\n \nPosenet\n \nare\n \nused\n \nto\n \nfollow\n \nobjects\n \bin\n \nframes\n \nin\n \norder\n \nto\n \nestimate\n \ntheir\n \npos es\n \nvia\n \nthe\n \nvideo\n \nfeed.\n \nIt\n \nwas\n \nobserved\n \nthat\n \ndetection\n \nwas\n \neffective\n \nand\n \nefficient.\n \nFor\n \nreal-time\n \ndetection\n \nand\n \nrecognition\n \napplications,\n \tthe\n \nalgorithms\n \nprovide\n \nreal-time,\n \naccurate,\n \nexact\n \nidentifications.\n \nAccording\n \ntest\n \nfindings,\n \nthe\n \nsystem\n \ncan\n \ncollect\n \nan\n \naverage\n \nof\n \n10.7\n \nframes\n \nper\n \nsecond\n \nand\n \nhas\n \na\n \ncamera\n \nmovement\n \nreaction\n \ntime\n \nof\n \nless\n \nthan\n \none\n \nsecond\n \nwhen\n \nthere\n \nare\n \nno\n \nmore\n \nthan\n \n2\n \nclients\n \nin\n \na\n \nnetwork.\n \nThe\n \nmonitoring\n \nsystem\n \becomes\n \na\n \nclever\n \nand\n \ndependable\n \nsystem\n \nafter\n \nsuccessful\n \ntesting\n \nof\n \nthe\n \nrecording\n \nand\n \nmotion\n \ndetecting\n \nfeatures.\n \nKeywords: Artificial Intelligence; Computer Vision; Convolution Neural \n \nNetwork;Single\n \nShot\n \nDetection;\n \nMobileNet;\n \nObject\n \ndetection;\n \nRasberry-Pi.\n \n \n \n6 ", + "attributes": [ + { + "attribute": "section", + "value": "Abstract" + }, + { + "attribute": "section", + "value": "Abstract" + } + ] + }, + { + "type": "fact", + "insight": "The CNN model is intended for object detection on the COCO dataset.", + "content": "ABSTRACT In the modern technological civilization, data is the new oil. The influence of \nefficient\n \n data\n has\n \naltered\n \nperformance\n standards\n \nin\nterms\n \nof\n \naccuracy\n \nand\n \nspeed.\n \nEvery\n \nhospital\n \nand\n \nhousehold\n \nis\n \nhave\n \na\n \nmonitoring\n \nsystem\n \nas\n \nit\n \nis\n \nso\n \ncrucial.\n \nAs\n \na\n \nRaspberry\n \nPi\n \nis\n \nused\n \nin\n \nthe\n \nsystem's\n \narchitecture\n \nas\n \the\n \nmicrocontroller,\n \nwhile\n \ta\n \nmobile\n \napplication\n \nis\n \ndeveloped\n \nfor\n \nmonitoring\n \nactivities\n \ndetected\n \nin\n \nthe\n \nsystem\n \nto\n \nalert\n \tthe\n \nuser\n \ntogether\n \nwith\n \na\n \nuser-controlled\n \nvideo\n \nrecording\n \nsystem.\n \nThe\n \nimprovement\n \nmay\n \nbe\n \nseen\n \nsince\n \ndata\n \nprocessing\n \nwas\n \nhandled\n \nby\n \ntwo\n \ncurrent\n \nindustry\n \nbuzzwords,\n \nComputer\n \nVision\n \n(CV)\n \nand\n \nArtificial\n \nIntelligence\n \n(AI).\n \nTwo\n \ntechnologies\n \nhave\n \nmade\n \nit\n \npossible\n \nto\n \ndo\n \nimportant\n \njobs\n \nlike\n \nobject\n \ndetecting\n \nsystems.\n \nAs\n \nthe\n \nnumber\n \nof\n \nfeatures\n \nin\n \na\n \npicture\n \ngrows,\n \nso\n \ndoes\n \nthe\n \nnecessity\n \nfor\n \nan\n \neffective\n \ttechnique\n \nto\n \nextract\n \nhidden\n \ninformation.\n \nThe\n \nCNN\n \nmodel\n \nis\n \nintended\n \nfor\n \nobject\n \ndetection\n \non\n \nthe\n \nCOCO\n \ndataset.\n \nAccuracy\n \nand\n \nprecision\n \nare\n \ntwo\n \nperformance\n \nmeasures\n \nthat\n \nare\n \nused\n \nto\n \nevaluate\n \nand\n \nverify\n \nthe\n \nmodel's\n \nperformance.\n \nSingle\n \nShot\n \nDetection\n \n(SSD),\n \nmobileNet,\n \nand\n \nPosenet\n \nare\n \nused\n \nto\n \nfollow\n \nobjects\n \bin\n \nframes\n \nin\n \norder\n \nto\n \nestimate\n \ntheir\n \npos es\n \nvia\n \nthe\n \nvideo\n \nfeed.\n \nIt\n \nwas\n \nobserved\n \nthat\n \ndetection\n \nwas\n \neffective\n \nand\n \nefficient.\n \nFor\n \nreal-time\n \ndetection\n \nand\n \nrecognition\n \napplications,\n \tthe\n \nalgorithms\n \nprovide\n \nreal-time,\n \naccurate,\n \nexact\n \nidentifications.\n \nAccording\n \ntest\n \nfindings,\n \nthe\n \nsystem\n \ncan\n \ncollect\n \nan\n \naverage\n \nof\n \n10.7\n \nframes\n \nper\n \nsecond\n \nand\n \nhas\n \na\n \ncamera\n \nmovement\n \nreaction\n \ntime\n \nof\n \nless\n \nthan\n \none\n \nsecond\n \nwhen\n \nthere\n \nare\n \nno\n \nmore\n \nthan\n \n2\n \nclients\n \nin\n \na\n \nnetwork.\n \nThe\n \nmonitoring\n \nsystem\n \becomes\n \na\n \nclever\n \nand\n \ndependable\n \nsystem\n \nafter\n \nsuccessful\n \ntesting\n \nof\n \nthe\n \nrecording\n \nand\n \nmotion\n \ndetecting\n \nfeatures.\n \nKeywords: Artificial Intelligence; Computer Vision; Convolution Neural \n \nNetwork;Single\n \nShot\n \nDetection;\n \nMobileNet;\n \nObject\n \ndetection;\n \nRasberry-Pi.\n \n \n \n6 ", + "attributes": [ + { + "attribute": "section", + "value": "Abstract" + }, + { + "attribute": "model", + "value": "CNN for object detection on COCO" + } + ] + }, + { + "type": "fact", + "insight": "SSD, MobileNet, and PoseNet are used to follow objects between frames to estimate their poses via the video feed.", + "content": "ABSTRACT In the modern technological civilization, data is the new oil. The influence of \nefficient\n \n data\n has\n \naltered\n \nperformance\n standards\n \nin\nterms\n \nof\n \naccuracy\n \nand\n \nspeed.\n \nEvery\n \nhospital\n \nand\n \nhousehold\n \nis\n \nhave\n \na\n \nmonitoring\n \nsystem\n \nas\n \tit\n \nis\n \nso\n \ncrucial.\n \nAs\n \na\n \nRaspberry\n \nPi\n \nis\n \nused\n \nin\n \nthe\n \nsystem's\n \narchitecture\n \nas\n \the\n \nmicrocontroller,\n \nwhile\n \ta\n \nmobile\n \napplication\n \nis\n \ndeveloped\n \nfor\n \nmonitoring\n \nactivities\n \ndetected\n \nin\n \nthe\n \nsystem\n \nto\n \nalert\n \tthe\n \nuser\n \ntogether\n \nwith\n \na\n \nuser-controlled\n \nvideo\n \nrecording\n \nsystem.\n \nThe\n \nimprovement\n \nmay\n \nbe\n \nseen\n \nsince\n \ndata\n \nprocessing\n \nwas\n \nhandled\n \nby\n \ntwo\n \ncurrent\n \nindustry\n \nbuzzwords,\n \nComputer\n \nVision\n \n(CV)\n \nand\n \nArtificial\n \nIntelligence\n \n(AI).\n \nTwo\n \ntechnologies\n \nhave\n \nmade\n \nit\n \npossible\n \nto\n \ndo\n \nimportant\n \njobs\n \nlike\n \nobject\n \ndetecting\n \nsystems.\n \nAs\n \nthe\n \nnumber\n \nof\n \nfeatures\n \nin\n \na\n \npicture\n \ngrows,\n \nso\n \ndoes\n \nthe\n \nnecessity\n \nfor\n \nan\n \neffective\n \ttechnique\n \nto\n \nextract\n \nhidden\n \ninformation.\n \nThe\n \nCNN\n \nmodel\n \nis\n \nintended\n \nfor\n \nobject\n \ndetection\n \non\n \nthe\n \nCOCO\n \ndataset.\n \nAccuracy\n \nand\n \nprecision\n \nare\n \ntwo\n \nperformance\n \nmeasures\n \nthat\n \nare\n \nused\n \nto\n \nevaluate\n \nand\n \nverify\n \nthe\n \nmodel's\n \nperformance.\n \nSingle\n \nShot\n \nDetection\n \n(SSD),\n \nmobileNet,\n \nand\n \nPosenet\n \nare\n \nused\n \nto\n \nfollow\n \nobjects\n \bin\n \nframes\n \nin\n \norder\n \nto\n \nestimate\n \ntheir\n \npos es\n \nvia\n \nthe\n \nvideo\n \nfeed.\n \nIt\n \nwas\n \nobserved\n \nthat\n \ndetection\n \nwas\n \neffective\n \nand\n \nefficient.\n \nFor\n \nreal-time\n \ndetection\n \nand\n \nrecognition\n \napplications,\n \tthe\n \nalgorithms\n \nprovide\n \nreal-time,\n \naccurate,\n \nexact\n \nidentifications.\n \nAccording\n \ntest\n \nfindings,\n \nthe\n \nsystem\n \ncan\n \ncollect\n \nan\n \naverage\n \nof\n \n10.7\n \nframes\n \nper\n \nsecond\n \nand\n \nhas\n \na\n \ncamera\n \nmovement\n \nreaction\n \ntime\n \nof\n \nless\n \nthan\n \none\n \nsecond\n \nwhen\n \nthere\n \nare\n \nno\n \nmore\n \nthan\n \n2\n \nclients\n \nin\n \na\n \nnetwork.\n \nThe\n \nmonitoring\n \nsystem\n \becomes\n \na\n \nclever\n \nan d\n \ndependable\n \nsystem\n \nafter\n \nsuccessful\n \ntesting\n \nof\n \nthe\n \nrecording\n \nand\n \nmotion\n \ndetecting\n \nfeatures.\n \nKeywords: Artificial Intelligence; Computer Vision; Convolution Neural \n \nNetwork;Single\n \nShot\n \nDetection;\n \nMobileNet;\n \nObject\n \ndetection;\n \nRasberry-Pi.\n \n \n \n6 ", + "attributes": [ + { + "attribute": "section", + "value": "Abstract" + }, + { + "attribute": "models", + "value": "SSD, MobileNet, PoseNet" + }, + { + "attribute": "task", + "value": "object tracking/pose estimation" + } + ] + }, + { + "type": "fact", + "insight": "Camera movement reaction time is less than one second under network conditions with two or fewer clients.", + "content": "ABSTRACT In the modern technological civilization, data is the new oil. The influence of \nefficient\n \n data\n has\n \naltered\n \nperformance\n standards\n \nin\nterms\n \nof\n \naccuracy\n \nand\n \nspeed.\n \nEvery\n \nhospital\n \nand\n \nhousehold\n \nis\n \nhave\n \na\n \nmonitoring\n \nsystem\n \nas\n \tit\n \nis\n \nso\n \ncrucial.\n \nAs\n \na\n \nRaspberry\n \nPi\n \nis\n \nused\n \nin\n \nthe\n \nsystem's\n \narchitecture\n \nas\n \the\n \nmicrocontroller,\n \nwhile\n \ta\n \nmobile\n \napplication\n \nis\n \ndeveloped\n \nfor\n \nmonitoring\n \nactivities\n \ndetected\n \nin\n \nthe\n \nsystem\n \nto\n \nalert\n \tthe\n \nuser\n \ntogether\n \nwith\n \na\n \nuser-controlled\n \nvideo\n \nrecording\n \nsystem.\n \nThe\n \nimprovement\n \nmay\n \nbe\n \nseen\n \nsince\n \ndata\n \nprocessing\n \nwas\n \nhandled\n \nby\n \ntwo\n \ncurrent\n \nindustry\n \nbuzzwords,\n \nComputer\n \nVision\n \n(CV)\n \nand\n \nArtificial\n \nIntelligence\n \n(AI).\n \nTwo\n \ntechnologies\n \nhave\n \nmade\n \nit\n \npossible\n \nto\n \ndo\n \nimportant\n \njobs\n \nlike\n \nobject\n \ndetecting\n \nsystems.\n \nAs\n \nthe\n \nnumber\n \nof\n \nfeatures\n \nin\n \na\n \npicture\n \ngrows,\n \nso\n \ndoes\n \nthe\n \nnecessity\n \nfor\n \nan\n \neffective\n \ttechnique\n \nto\n \nextract\n \nhidden\n \ninformation.\n \nThe\n \nCNN\n \nmodel\n \nis\n \nintended\n \nfor\n \nobject\n \ndetection\n \non\n \nthe\n \nCOCO\n \ndataset.\n \nAccuracy\n \nand\n \nprecision\n \nare\n \ntwo\n \nperformance\n \nmeasures\n \nthat\n \nare\n \nused\n \nto\n \nevaluate\n \nand\n \nverify\n \nthe\n \nmodel's\n \nperformance.\n \nSingle\n \nShot\n \nDetection\n \n(SSD),\n \nmobileNet,\n \nand\n \nPosenet\n \nare\n \nused\n \nto\n \nfollow\n \nobjects\n \bin\n \nframes\n \nin\n \norder\n \nto\n \nestimate\n \ntheir\n \npos es\n \nvia\n \nthe\n \nvideo\n \nfeed.\n \nIt\n \nwas\n \nobserved\n \nthat\n \ndetection\n \nwas\n \neffective\n \nand\n \nefficient.\n \nFor\n \nreal-time\n \ndetection\n \nand\n \nrecognition\n \napplications,\n \tthe\n \nalgorithms\n \nprovide\n \nreal-time,\n \naccurate,\n \nexact\n \nidentifications.\n \nAccording\n \ntest\n \nfindings,\n \nthe\n \nsystem\n \ncan\n \ncollect\n \nan\n \naverage\n \nof\n \n10.7\n \nframes\n \nper\n \nsecond\n \nand\n \nhas\n \na\n \ncamera\n \nmovement\n \nreaction\n \ntime\n \nof\n \nless\n \nthan\n \none\n \nsecond\n \nwhen\n \nthere\n \nare\n \nno\n \nmore\n \nthan\n \n2\n \nclients\n \nin\n \na\n \nnetwork.\n \nThe\n \nmonitoring\n \nsystem\n \becomes\n \na\n \nclever\n \nan d\n \ndependable\n \nsystem\n \nafter\n \nsuccessful\n \ntesting\n \nof\n \nthe\n \nrecording\n \nand\n \nmotion\n \ndetecting\n \nfeatures.\n \nKeywords: Artificial Intelligence; Computer Vision; Convolution Neural \n \nNetwork;Single\n \nShot\n \nDetection;\n \nMobileNet;\n \nObject\n \ndetection;\n \nRasberry-Pi.\n \n \n \n6 ", + "attributes": [ + { + "attribute": "section", + "value": "Abstract" + }, + { + "attribute": "section", + "value": "Abstract" + } + ] + }, + { + "type": "comment", + "insight": "There is a misspelling in the keywords: Raspberry Pi is spelled Rasberry-Pi.", + "content": "ABSTRACT In the modern technological civilization, data is the new oil. The influence of \nefficient\n \n data\n has\n \naltered\n \nperformance\n standards\n \nin\nterms\n \nof\n \naccuracy\n \nand\n \nspeed.\n \nEvery\n \nhospital\n \nand\n \nhousehold\n \nis\n \nhave\n \na\n \nmonitoring\n \nsystem\n \nas\n \tit\n \nis\n \nso\n \ncrucial.\n \nAs\n \na\n \nRaspberry\n \nPi\n \nis\n \nused\n \nin\n \nthe\n \nsystem's\n \narchitecture\n \nas\n \the\n \nmicrocontroller,\n \nwhile\n \ta\n \nmobile\n \napplication\n \nis\n \ndeveloped\n \nfor\n \nmonitoring\n \nactivities\n \ndetected\n \nin\n \nthe\n \nsystem\n \nto\n \nalert\n \tthe\n \nuser\n \ntogether\n \nwith\n \na\n \nuser-controlled\n \nvideo\n \nrecording\n \nsystem.\n \nThe\n \nimprovement\n \nmay\n \nbe\n \nseen\n \nsince\n \ndata\n \nprocessing\n \nwas\n \nhandled\n \nby\n \ntwo\n \ncurrent\n \nindustry\n \nbuzzwords,\n \nComputer\n \nVision\n \n(CV)\n \nand\n \nArtificial\n \nIntelligence\n \n(AI).\n \nTwo\n \ntechnologies\n \nhave\n \nmade\n \nit\n \npossible\n \nto\n \ndo\n \nimportant\n \njobs\n \nlike\n \nobject\n \ndetecting\n \nsystems.\n \nAs\n \nthe\n \nnumber\n \nof\n \nfeatures\n \nin\n \na\n \npicture\n \ngrows,\n \nso\n \ndoes\n \nthe\n \nnecessity\n \nfor\n \nan\n \neffective\n \ttechnique\n \nto\n \nextract\n \nhidden\n \ninformation.\n \nThe\n \nCNN\n \nmodel\n \nis\n \nintended\n \nfor\n \nobject\n \ndetection\n \non\n \nthe\n \nCOCO\n \ndataset.\n \nAccuracy\n \nand\n \nprecision\n \nare\n \ntwo\n \nperformance\n \nmeasures\n \nthat\n \nare\n \nused\n \nto\n \nevaluate\n \nand\n \nverify\n \nthe\n \nmodel's\n \nperformance.\n \nSingle\n \nShot\n \nDetection\n \n(SSD),\n \nmobileNet,\n \nand\n \nPosenet\n \nare\n \nused\n \nto\n \nfollow\n \nobjects\n \bin\n \nframes\n \nin\n \norder\n \nto\n \nestimate\n \ntheir\n \npos es\n \nvia\n \nthe\n \nvideo\n \nfeed.\n \nIt\n \nwas\n \nobserved\n \nthat\n \ndetection\n \nwas\n \neffective\n \nand\n \nefficient.\n \nFor\n \nreal-time\n \ndetection\n \nand\n \nrecognition\n \napplications,\n \tthe\n \nalgorithms\n \nprovide\n \nreal-time,\n \naccurate,\n \nexact\n \nidentifications.\n \nAccording\n \ntest\n \nfindings,\n \nthe\n \nsystem\n \ncan\n \ncollect\n \nan\n \naverage\n \nof\n \n10.7\n \nframes\n \nper\n \nsecond\n \nand\n \nhas\n \na\n \ncamera\n \nmovement\n \nreaction\n \ntime\n \nof\n \nless\n \nthan\n \none\n \nsecond\n \nwhen\n \nthere\n \nare\n \nno\n \nmore\n \nthan\n \n2\n \nclients\n \nin\n \na\n \nnetwork.\n \nThe\n \nmonitoring\n \nsystem\n \becomes\n \na\n \nclever\n \nan d\n \ndependable\n \nsystem\n \nafter\n \nsuccessful\n \ntesting\n \nof\n \nthe\n \nrecording\n \nand\n \nmotion\n \ndetecting\n \nfeatures.\n \nKeywords: Artificial Intelligence; Computer Vision; Convolution Neural \n \nNetwork;Single\n \nShot\n \nDetection;\n \nMobileNet;\n \nObject\n \ndetection;\n \nRasberry-Pi.\n \n \n \n6 ", + "attributes": [ + { + "attribute": "section", + "value": "Abstract" + }, + { + "attribute": "note", + "value": "misspelling Raspberry Pi as Rasberry-Pi" + } + ] + }, + { + "type": "fact", + "insight": "Mobile application is developed for monitoring and alerts, with a user-controlled video recording system.", + "content": "ABSTRACT In the modern technological civilization, data is the new oil. The influence of \nefficient\n \n data\n has\n \naltered\n \nperformance\n standards\n \nin\nterms\n \nof\n \naccuracy\n \nand\n \nspeed.\n \nEvery\n \nhospital\n \nand\n \nhousehold\n \nis\n \nhave\n \na\n \nmonitoring\n \nsystem\n \nas\n \tit\n \nis\n \nso\n \ncrucial.\n \nAs\n \na\n \nRaspberry\n \nPi\n \nis\n \nused\n \nin\n \nthe\n \nsystem's\n \narchitecture\n \nas\n \the\n \nmicrocontroller,\n \nwhile\n \ta\n \nmobile\n \napplication\n \nis\n \ndeveloped\n \nfor\n \nmonitoring\n \nactivities\n \ndetected\n \nin\n \nthe\n \nsystem\n \nto\n \nalert\n \tthe\n \nuser\n \ntogether\n \nwith\n \na\n \nuser-controlled\n \nvideo\n \nrecording\n \nsystem.\n \nThe\n \nimprovement\n \nmay\n \nbe\n \nseen\n \nsince\n \ndata\n \nprocessing\n \nwas\n \nhandled\n \nby\n \ntwo\n \ncurrent\n \nindustry\n \nbuzzwords,\n \nComputer\n \nVision\n \n(CV)\n \nand\n \nArtificial\n \nIntelligence\n \n(AI).\n \nTwo\n \ntechnologies\n \nhave\n \nmade\n \nit\n \npossible\n \nto\n \ndo\n \nimportant\n \njobs\n \nlike\n \nobject\n \ndetecting\n \nsystems.\n \nAs\n \nthe\n \nnumber\n \nof\n \nfeatures\n \nin\n \na\n \npicture\n \ngrows,\n \nso\n \ndoes\n \nthe\n \nnecessity\n \nfor\n \nan\n \neffective\n \ttechnique\n \nto\n \nextract\n \nhidden\n \ninformation.\n \nThe\n \nCNN\n \nmodel\n \nis\n \nintended\n \nfor\n \nobject\n \ndetection\n \non\n \nthe\n \nCOCO\n \ndataset.\n \nAccuracy\n \nand\n \nprecision\n \nare\n \ntwo\n \nperformance\n \nmeasures\n \nthat\n \nare\n \nused\n \nto\n \nevaluate\n \nand\n \nverify\n \nthe\n \nmodel's\n \nperformance.\n \nSingle\n \nShot\n \nDetection\n \n(SSD),\n \nmobileNet,\n \nand\n \nPosenet\n \nare\n \nused\n \nto\n \nfollow\n \nobjects\n \bin\n \nframes\n \nin\n \norder\n \nto\n \nestimate\n \ntheir\n \npos es\n \nvia\n \nthe\n \nvideo\n \nfeed.\n \nIt\n \nwas\n \nobserved\n \nthat\n \ndetection\n \nwas\n \neffective\n \nand\n \nefficient.\n \nFor\n \nreal-time\n \ndetection\n \nand\n \nrecognition\n \napplications,\n \tthe\n \nalgorithms\n \nprovide\n \nreal-time,\n \naccurate,\n \nexact\n \nidentifications.\n \nAccording\n \ntest\n \nfindings,\n \nthe\n \nsystem\n \ncan\n \ncollect\n \nan\n \naverage\n \nof\n \n10.7\n \nframes\n \nper\n \nsecond\n \nand\n \nhas\n \na\n \ncamera\n \nmovement\n \nreaction\n \ntime\n \nof\n \nless\n \nthan\n \none\n \nsecond\n \nwhen\n \nthere\n \nare\n \nno\n \nmore\n \nthan\n \n2\n \nclients\n \nin\n \na\n \nnetwork.\n \nThe\n \nmonitoring\n \nsystem\n \becomes\n \na\n \nclever\n \nan d\n \ndependable\n \nsystem\n \nafter\n \nsuccessful\n \ntesting\n \nof\n \nthe\n \nrecording\n \nand\n \nmotion\n \ndetecting\n \nfeatures.\n \nKeywords: Artificial Intelligence; Computer Vision; Convolution Neural \n \nNetwork;Single\n \nShot\n \nDetection;\n \nMobileNet;\n \nObject\n \ndetection;\n \nRasberry-Pi.\n \n \n \n6 ", + "attributes": [ + { + "attribute": "section", + "value": "Abstract" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "CNN model is intended for object detection on the COCO dataset.", + "content": "ABSTRACT In the modern technological civilization, data is the new oil. The influence of \nefficient\n \n data\n has\n \naltered\n \nperformance\n standards\n \nin\nterms\n \nof\n \naccuracy\n \nand\n \nspeed.\n \nEvery\n \nhospital\n \nand\n \nhousehold\n \nis\n \nhave\n \na\n \nmonitoring\n \nsystem\n \nas\n \nit\n \nis\n \nso\n \ncrucial.\n \nAs\n \na\n \nRaspberry\n \nPi\n \nis\n \nused\n \nin\n \nthe\n \nsystem's\n \narchitecture\n \nas\n \the\n \nmicrocontroller,\n \nwhile\n \ta\n \nmobile\n \napplication\n \nis\n \ndeveloped\n \nfor\n \nmonitoring\n \nactivities\n \ndetected\n \nin\n \nthe\n \nsystem\n \nto\n \nalert\n \tthe\n \nuser\n \ntogether\n \nwith\n \na\n \nuser-controlled\n \nvideo\n \nrecording\n \nsystem.\n \nThe\n \nimprovement\n \nmay\n \nbe\n \nseen\n \nsince\n \ndata\n \nprocessing\n \nwas\n \nhandled\n \nby\n \ntwo\n \ncurrent\n \nindustry\n \nbuzzwords,\n \nComputer\n \nVision\n \n(CV)\n \nand\n \nArtificial\n \nIntelligence\n \n(AI).\n \nTwo\n \ntechnologies\n \nhave\n \nmade\n \nit\n \npossible\n \nto\n \ndo\n \nimportant\n \njobs\n \nlike\n \nobject\n \ndetecting\n \nsystems.\n \nAs\n \nthe\n \nnumber\n \nof\n \nfeatures\n \nin\n \na\n \npicture\n \ngrows,\n \nso\n \ndoes\n \nthe\n \nnecessity\n \nfor\n \nan\n \neffective\n \ttechnique\n \nto\n \nextract\n \nhidden\n \ninformation.\n \nThe\n \nCNN\n \nmodel\n \nis\n \nintended\n \nfor\n \nobject\n \ndetection\n \non\n \nthe\n \nCOCO\n \ndataset.\n \nAccuracy\n \nand\n \nprecision\n \nare\n \ntwo\n \nperformance\n \nmeasures\n \nthat\n \nare\n \nused\n \nto\n \nevaluate\n \nan d\n \nverify\n \nthe\n \nmodel's\n \nperformance.\n \nSingle\n \nShot\n \nDetection\n \n(SSD),\n \nmobileNet,\n \nand\n \nPosenet\n \nare\n \nused\n \nto\n \nfollow\n \nobjects\n \bin\n \nframes\n \nin\n \norder\n \nto\n \nestimate\n \ntheir\n \npos es\n \nvia\n \nthe\n \nvideo\n \nfeed.\n \nIt\n \nwas\n \nobserved\n \ndetection\n \nwas\n \neffective\n \nand\n \nefficient.\n \nFor\n \nreal-time\n \ndetection\n \nand\n \nrecognition\n \napplications,\n \tthe\n \nalgorithms\n \nprovide\n \nreal-time,\n \naccurate,\n \nexact\n \nidentifications.\n \nAccording\n \ntest\n \nfindings,\n \nthe\n \nsystem\n \ncan\n \ncollect\n \nan\n \naverage\n \nof\n \n10.7\n \nframes\n \nper\n \nsecond\n \nand\n \nhas\n \na\n \ncamera\n \nmovement\n \nreaction\n \ntime\n \nof\n \nless\n \nthan\n \none\n \nsecond\n \nwhen\n \nthere\n \nare\n \nno\n \nmore\n \nthan\n \n2\n \nclients\n \nin\n \na\n \nnetwork.\n \nThe\n \nmonitoring\n \nsystem\n \becomes\n \na\n \nclever\n \nand\n \ndependable\n \nsystem\n \nafter\n \nsuccessful\n \ntesting\n \nof\n \nthe\n \nrecording\n \nand\n \nmotion\n \ndetecting\n \nfeatures.\n \nKeywords: Artificial Intelligence; Computer Vision; Convolution Neural \n \nNetwork;Single\n \nShot\n \nDetection;\n \nMobileNet;\n \nObject\n \ndetection;\n \nRasberry-Pi.\n \n \n \n6 ", + "attributes": [ + { + "attribute": "section", + "value": "Abstract" + }, + { + "attribute": "attribute", + "value": "Raspberry Pi misspelled as Rasberry-Pi" + } + ] + }, + { + "type": "fact", + "insight": "The mobile app and video recording feature are described, with test findings.", + "content": "ABSTRACT In the modern technological civilization, data is the new oil. The influence of \nefficient\n \n data\n has\n \naltered\n \nperformance\n standards\n \nin\nterms\n \nof\n \naccuracy\n \nand\n \nspeed.\n \nEvery\n \nhospital\n \nand\n \nhousehold\n \nis\n \nhave\n \na\n \nmonitoring\n \nsystem\n \nas\n \tit\n \nis\n \nso\n \ncrucial.\n \nAs\n \na\n \nRaspberry\n \nPi\n \nis\n \nused\n \nin\n \nthe\n \nsystem's\n \narchitecture\n \nas\n \the\n \nmicrocontroller,\n \nwhile\n \ta\n \nmobile\n \napplication\n \nis\n \ndeveloped\n \nfor\n \nmonitoring\n \nactivities\n \ndetected\n \nin\n \nthe\n \nsystem\n \nto\n \nalert\n \tthe\n \nuser\n \ntogether\n \nwith\n \na\n \nuser-controlled\n \nvideo\n \nrecording\n \nsystem.\n \nThe\n \nimprovement\n \nmay\n \nbe\n \nseen\n \nsince\n \ndata\n \nprocessing\n \nwas\n \nhandled\n \nby\n \ntwo\n \ncurrent\n \nindustry\n \nbuzzwords,\n \nComputer\n \nVision\n \n(CV)\n \nand\n \nArtificial\n \nIntelligence\n \n(AI).\n \nTwo\n \ntechnologies\n \nhave\n \nmade\n \nit\n \npossible\n \nto\n \ndo\n \nimportant\n \njobs\n \nlike\n \nobject\n \ndetecting\n \nsystems.\n \nAs\n \nthe\n \nnumber\n \nof\n \nfeatures\n \nin\n \na\n \npicture\n \ngrows,\n \nso\n \ndoes\n \nthe\n \nnecessity\n \nfor\n \nan\n \neffective\n \ttechnique\n \nto\n \nextract\n \nhidden\n \ninformation.\n \nThe\n \nCNN\n \nmodel\n \nis\n \nintended\n \nfor\n \nobject\n \ndetection\n \non\n \nthe\n \nCOCO\n \ndataset.\n \nAccuracy\n \nand\n \nprecision\n \nare\n \ntwo\n \nperformance\n \nmeasures\n \nthat\n \nare\n \nused\n \nto\n \nevaluate\n \nand\n \nverify\n \nthe\n \nmodel's\n \nperformance.\n \nSingle\n \nShot\n \nDetection\n \n(SSD),\n \nmobileNet,\n \nand\n \nPosenet\n \nare\n \nused\n \nto\n \nfollow\n \nobjects\n \bin\n \nframes\n \nin\n \norder\n \nto\n \nestimate\n \ntheir\n \npos es\n \nvia\n \nthe\n \nvideo\n \nfeed.\n \nIt\n \nwas\n \nobserved\n \ndetection\n \nwas\n \neffective\n \nand\n \nefficient.\n \nFor\n \nreal-time\n \ndetection\n \nand\n \nrecognition\n \napplications,\n \tthe\n \nalgorithms\n \nprovide\n \nreal-time,\n \naccurate,\n \nexact\n \nidentifications.\n \nAccording\n \ntest\n \nfindings,\n \nthe\n \nsystem\n \ncan\n \ncollect\n \nan\n \naverage\n \nof\n \n10.7\n \nframes\n \nper\n \nsecond\n \nand\n \nhas\n \na\n \ncamera\n \nmovement\n \nreaction\n \ntime\n \nof\n \nless\n \nthan\n \none\n \nsecond\n \nwhen\n \nthere\n \nare\n \nno\n \nmore\n \nthan\n \n2\n \nclients\n \nin\n \na\n \nnetwork.\n \nThe\n \nmonitoring\n \nsystem\n \becomes\n \na\n \nclever\n \nand\n \ndependable\n \nsystem\n \nafter\n \nsuccessful\n \ntesting\n \nof\n \nthe\n \nrecording\n \nand\n \nmotion\n \ndetecting\n \nfeatures.\n \nKeywords: Artificial Intelligence; Computer Vision; Convolution Neural \n \nNetwork;Single\n \nShot\n \nDetection;\n \nMobileNet;\n \nObject\n \ndetection;\n \nRasberry-Pi.\n \n \n \n6 ", + "attributes": [ + { + "attribute": "section", + "value": "Abstract" + }, + { + "attribute": "note", + "value": "additional context" + } + ] + }, + { + "type": "fact", + "insight": "The abstract mentions Raspberry Pi spelling as Rasberry-Pi in keywords.", + "content": "ABSTRACT In the modern technological civilization, data is the new oil. The influence of \nefficient\n \n data\n has\n \naltered\n \nperformance\n standards\n \nin\nterms\n \nof\n \naccuracy\n \nand\n \nspeed.\n \nEvery\n \nhospital\n \nand\n \nhousehold\n \nis\n \nhave\n \na\n \nmonitoring\n \nsystem\n \nas\n \tit\n \nis\n \nso\n \ncrucial.\n \nAs\n \na\n \nRaspberry\n \nPi\n \nis\n \nused\n \nin\n \nthe\n \nsystem's\n \narchitecture\n \nas\n \the\n \nmicrocontroller,\n \nwhile\n \ta\n \nmobile\n \napplication\n \nis\n \ndeveloped\n \nfor\n \nmonitoring\n \nactivities\n \ndetected\n \nin\n \nthe\n \nsystem\n \nto\n \nalert\n \tthe\n \nuser\n \ntogether\n \nwith\n \na\n \nuser-controlled\n \nvideo\n \nrecording\n \nsystem.\n \nThe\n \nimprovement\n \nmay\n \nbe\n \nseen\n \nsince\n \ndata\n \nprocessing\n \nwas\n \nhandled\n \nby\n \ntwo\n \ncurrent\n \nindustry\n \nbuzzwords,\n \nComputer\n \nVision\n \n(CV)\n \nand\n \nArtificial\n \nIntelligence\n \n(AI).\n \nTwo\n \ntechnologies\n \nhave\n \nmade\n \nit\n \npossible\n \nto\n \ndo\n \nimportant\n \njobs\n \nlike\n \nobject\n \ndetecting\n \nsystems.\n \nAs\n \nthe\n \nnumber\n \nof\n \nfeatures\n \nin\n \na\n \npicture\n \ngrows,\n \nso\n \ndoes\n \nthe\n \nnecessity\n \nfor\n \nan\n \neffective\n \ttechnique\n \nto\n \nextract\n \nhidden\n \ninformation.\n \nThe\n \nCNN\n \nmodel\n \nis\n \nintended\n \nfor\n \nobject\n \ndetection\n \non\n \nthe\n \nCOCO\n \ndataset.\n \nAccuracy\n \nand\n \nprecision\n \nare\n \ntwo\n \nperformance\n \nmeasures\n \nthat\n \nare\n \nused\n \nto\n \nevaluate\n \nand\n \nverify\n \nthe\n \nmodel's\n \nperformance.\n \nSingle\n \nShot\n \nDetection\n \n(SSD),\n \nmobileNet,\n \nand\n \nPosenet\n \nare\n \nused\n \nto\n \nfollow\n \nobjects\n \bin\n \nframes\n \nin\n \norder\n \nto\n \nestimate\n \ntheir\n \npos es\n \nvia\n \nthe\n \nvideo\n \nfeed.\n \nIt\n \nwas\n \nobserved\n \ndetection\n \nwas\n \neffective\n \nand\n \nefficient.\n \nFor\n \nreal-time\n \ndetection\n \nand\n \nrecognition\n \napplications,\n \tthe\n \nalgorithms\n \nprovide\n \nreal-time,\n \naccurate,\n \nexact\n \nidentifications.\n \nAccording\n \ntest\n \nfindings,\n \nthe\n \nsystem\n \ncan\n \ncollect\n \nan\n \naverage\n \nof\n \n10.7\n \nframes\n \nper\n \nsecond\n \nand\n \nhas\n \na\n \ncamera\n \nmovement\n \nreaction\n \ntime\n \nof\n \nless\n \nthan\n \none\n \nsecond\n \nwhen\n \nthere\n \nare\n \nno\n \nmore\n \nthan\n \n2\n \nclients\n \nin\n \na\n \nnetwork.\n \nThe\n \nmonitoring\n \nsystem\n \becomes\n \na\n \nclever\n \nan d\n \ndependable\n \nsystem\n \nafter\n \nsuccessful\n \ntesting\n \nof\n \nthe\n \nrecording\n \nand\n \nmotion\n \ndetecting\n \nfeatures.\n \nKeywords: Artificial Intelligence; Computer Vision; Convolution Neural \n \nNetwork;Single\n \nShot\n \nDetection;\n \nMobileNet;\n \nObject\n \ndetection;\n \nRasberry-Pi.\n \n \n \n6 ", + "attributes": [ + { + "attribute": "section", + "value": "Abstract" + }, + { + "attribute": "note", + "value": "misspelling in keywords Rasberry-Pi" + } + ] + }, + { + "type": "fact", + "insight": "The page provides a detailed outline for Chapter 3 (Methodology), listing sections 3.0 Overview, 3.1 Overview, 3.2 Materials (software and hardware), 3.3 Method (design of a multi-function activity monitoring and reporting system; design of centralized database architecture; design of a mobile application), and 3.4 BEME (Bill of Engineering Measurement and Evaluation).", + "content": "2.6 Summary of the Reviewed Literatures ........................................................... 11 CHAPTER THREE .................................................................................................. 13 3.0 METHODOLOGY ............................................................................................. 13 3.1 Overview ........................................................................................................ 13 3.2 Materials ......................................................................................................... 13 3.2.1 Software Materials with description ........................................................ 13 3.2.2 Hardware Materials with description ...................................................... 15 3.3 Method ........................................................................................................... 16 3.3.1 Design of a multi-function activity monitoring and reporting system .... 17 3.3.2 Design of a centralized database architecture with storage scheme for multiple activity tracking and reporting system .......................................... 23 3.3.3 Design of a mobile application for managing monitored and reported activites ............................................................................................................. 23 3.4 Bill of Engineering Measurement and Evaluation (BEME) .......................... 25 Major Research Methods and its Findings ........................................................... 25 CHAPTER FOUR .................................................................................................... 26 4.0 RESULT AND DISCUSSION ............................................................................ 26 4.1 Overview ........................................................................................................ 26 4.2 Results and Discussions for the multi-function activity monitoring and reporting system ................................................................................................... 26 4.2.1 Motion Detection ..................................................................................... 27 4.2.2 Posture Recognition ................................................................................. 29 4.3 Results and Discussion for the centralized database architecture with storage scheme for multiple activity tracking and reporting system ................................ 29 4.3.1 Firebase Realtime database ..................................................................... 30 4.3.2Firebase Cloud Storage ............................................................................. 31 4.4 Results and Discussions for the mobile application for managing monitored and reported activites. .......................................................................................... 31 4.4.1 Introduction Screen .................................................................................. 32 4.4.2 Login Screen ............................................................................................ 32 4.4.3 Registration Screen .................................................................................. 33 4.4.4 Forget Password Screen ........................................................................... 34 4.4.5 Home Screen ............................................................................................ 35 4.4.6 Create Activity Screen ............................................................................. 36 4.4.7Activities Screen ....................................................................................... 38", + "attributes": [ + { + "attribute": "section", + "value": "Chapter 3 - Methodology (TOC)" + }, + { + "attribute": "page", + "value": "7" + }, + { + "attribute": "confidence", + "value": "0.85" + } + ] + }, + { + "type": "fact", + "insight": "The page lists Chapter Four (Results and Discussion) with subsections covering an overview, multi-function activity monitoring and reporting results (including Motion Detection and Posture Recognition), centralized database architecture results (Firebase Realtime Database and Firebase Cloud Storage), and mobile application results (UI screens such as Introduction, Login, Registration, Forget Password, Home, Create Activity, and Activities).", + "content": "4.0 RESULT AND DISCUSSION ............................................................................ 26 4.1 Overview ........................................................................................................ 26 4.2 Results and Discussions for the multi-function activity monitoring and reporting system ................................................................................................... 26 4.2.1 Motion Detection ..................................................................................... 27 4.2.2 Posture Recognition ................................................................................. 29 4.3 Results and Discussion for the centralized database architecture with storage scheme for multiple activity tracking and reporting system ................................ 29 4.3.1 Firebase Realtime database ..................................................................... 30 4.3.2Firebase Cloud Storage ............................................................................. 31 4.4 Results and Discussions for the mobile application for managing monitored and reported activites. .......................................................................................... 31 4.4.1 Introduction Screen .................................................................................. 32 4.4.2 Login Screen ............................................................................................ 32 4.4.3 Registration Screen .................................................................................. 33 4.4.4 Forget Password Screen ........................................................................... 34 4.4.5 Home Screen ............................................................................................ 35 4.4.6 Create Activity Screen ............................................................................. 36 4.4.7Activities Screen ....................................................................................... 38", + "attributes": [ + { + "attribute": "section", + "value": "Chapter 4 - Results and Discussion (TOC)" + }, + { + "attribute": "page", + "value": "7" + }, + { + "attribute": "confidence", + "value": "0.85" + } + ] + }, + { + "type": "comment", + "insight": "This page functions as a Table of Contents, outlining the structure and scope of the document rather than presenting experimental results on this page.", + "content": "2.6 Summary of the Reviewed Literatures ........................................................... 11 CHAPTER THREE .................................................................................................. 13 3.0 METHODOLOGY ............................................................................................. 13 3.1 Overview ........................................................................................................ 13 3.2 Materials ......................................................................................................... 13 3.2.1 Software Materials with description ........................................................ 13 3.2.2 Hardware Materials with description ...................................................... 15 3.3 Method ........................................................................................................... 16 3.3.1 Design of a multi-function activity monitoring and reporting system .... 17 3.3.2 Design of a centralized database architecture with storage scheme for multiple activity tracking and reporting system .......................................... 23 3.3.3 Design of a mobile application for managing monitored and reported activites ............................................................................................................. 23 3.4 Bill of Engineering Measurement and Evaluation (BEME) .......................... 25 Major Research Methods and its Findings ........................................................... 25 CHAPTER FOUR .................................................................................................... 26 4.0 RESULT AND DISCUSSION ............................................................................ 26 4.1 Overview ........................................................................................................ 26 4.2 Results and Discussions for the multi-function activity monitoring and reporting system ................................................................................................... 26 4.2.1 Motion Detection ..................................................................................... 27 4.2.2 Posture Recognition ................................................................................. 29 4.3 Results and Discussion for the centralized database architecture with storage scheme for multiple activity tracking and reporting system ................................ 29 4.3.1 Firebase Realtime database ..................................................................... 30 4.3.2Firebase Cloud Storage ............................................................................. 31 4.4 Results and Discussions for the mobile application for managing monitored and reported activites. .......................................................................................... 31 4.4.1 Introduction Screen .................................................................................. 32 4.4.2 Login Screen ............................................................................................ 32 4.4.3 Registration Screen .................................................................................. 33 4.4.4 Forget Password Screen ........................................................................... 34 4.4.5 Home Screen ............................................................................................ 35 4.4.6 Create Activity Screen ............................................................................. 36 4.4.7Activities Screen ....................................................................................... 38", + "attributes": [ + { + "attribute": "section", + "value": "TOC Copy" + }, + { + "attribute": "page", + "value": "7" + } + ] + }, + { + "type": "comment", + "insight": "Page 8 presents a structural table of contents for the document, including references to Chapter 2 (Summary of the Reviewed Literatures) and Chapter 3 (Methodology) followed by Chapter 4 (Result and Discussion).", + "content": "2.6 Summary of the Reviewed Literatures ........................................................... 11 CHAPTER THREE .................................................................................................. 13 3.0 METHODOLOGY ............................................................................................. 13 3.1 Overview ........................................................................................................ 13 3.2 Materials ......................................................................................................... 13 3.2.1 Software Materials with description ........................................................ 13 3.2.2 Hardware Materials with description ...................................................... 15 3.3 Method ........................................................................................................... 16 3.3.1 Design of a multi-function activity monitoring and reporting system .... 17 3.3.2 Design of a centralized database architecture with storage scheme for multiple activity tracking and reporting system .......................................... 23 3.3.3 Design of a mobile application for managing monitored and reported activites ............................................................................................................. 23 3.4 Bill of Engineering Measurement and Evaluation (BEME) .......................... 25 Major Research Methods and its Findings ........................................................... 25 CHAPTER FOUR .................................................................................................... 26 4.0 RESULT AND DISCUSSION ............................................................................ 26 4.1 Overview ........................................................................................................ 26 4.2 Results and Discussions for the multi-function activity monitoring and reporting system ................................................................................................... 26 4.2.1 Motion Detection ..................................................................................... 27 4.2.2 Posture Recognition ................................................................................. 29 4.3 Results and Discussion for centralized database architecture with storage scheme for multiple activity tracking and reporting system ................................ 29 4.3.1 Firebase Realtime database ..................................................................... 30 4.3.2Firebase Cloud Storage ............................................................................. 31 4.4 Results and Discussions for the mobile application for managing monitored and reported activites. .......................................................................................... 31 4.4.1 Introduction Screen .................................................................................. 32 4.4.2 Login Screen ............................................................................................ 32 4.4.3 Registration Screen .................................................................................. 33 4.4.4 Forget Password Screen ........................................................................... 34 4.4.5 Home Screen ............................................................................................ 35 4.4.6 Create Activity Screen ............................................................................. 36 4.4.7Activities Screen ...................................................................................... 38 \n", + "attributes": [ + { + "attribute": "section", + "value": "Page content/TOC for Chapters 2-4" + }, + { + "attribute": "page_number", + "value": "8" + }, + { + "attribute": "source", + "value": "Document TOC excerpt" + }, + { + "attribute": "confidence", + "value": "High" + }, + { + "attribute": "sentiment", + "value": "Neutral" + } + ] + }, + { + "type": "fact", + "insight": "Chapter 3 is titled METHODOLOGY and lists key subsections including 3.0 Overview, 3.2 Materials (Software and Hardware), 3.3 Method, and 3.4 BEME.", + "content": "CHAPTER THREE .................................................................................................. 13 3.0 METHODOLOGY ............................................................................................. 13 3.1 Overview ........................................................................................................ 13 3.2 Materials ......................................................................................................... 13 3.2.1 Software Materials with description ........................................................ 13 3.2.2 Hardware Materials with description ...................................................... 15 3.3 Method ........................................................................................................... 16 3.3.1 Design of a multi-function activity monitoring and reporting system .... 17 3.3.2 Design of a centralized database architecture with storage scheme for multiple activity tracking and reporting system .......................................... 23 3.3.3 Design of a mobile application for managing monitored and reported activites ............................................................................................................. 23 3.4 Bill of Engineering Measurement and Evaluation (BEME) .......................... 25 Major Research Methods and its Findings ........................................................... 25", + "attributes": [ + { + "attribute": "section", + "value": "Chapter 3 – Methodology (TOC)" + }, + { + "attribute": "page_number", + "value": "8" + }, + { + "attribute": "source", + "value": "Document TOC excerpt" + }, + { + "attribute": "confidence", + "value": "High" + }, + { + "attribute": "sentiment", + "value": "Neutral" + } + ] + }, + { + "type": "fact", + "insight": "Chapter 4 is titled RESULT AND DISCUSSION and contains subsections such as 4.0 Overview, 4.2 Results for the multi-function activity monitoring and reporting system (with 4.2.1 Motion Detection and 4.2.2 Posture Recognition), 4.3 Results for the centralized database architecture (with 4.3.1 Firebase Realtime database and 4.3.2 Firebase Cloud Storage), and 4.4 Results for the mobile application (with 4.4.1 Introduction Screen through 4.4.7 Activities Screen).", + "content": "CHAPTER FOUR .................................................................................................... 26 4.0 RESULT AND DISCUSSION ............................................................................ 26 4.1 Overview ........................................................................................................ 26 4.2 Results and Discussions for the multi-function activity monitoring and reporting system ................................................................................................... 26 4.2.1 Motion Detection ..................................................................................... 27 4.2.2 Posture Recognition ................................................................................. 29 4.3 Results and Discussion for centralized database architecture with storage scheme for multiple activity tracking and reporting system ................................ 29 4.3.1 Firebase Realtime database ..................................................................... 30 4.3.2Firebase Cloud Storage ............................................................................. 31 4.4 Results and Discussions for the mobile application for managing monitored and reported activites. .......................................................................................... 31 4.4.1 Introduction Screen .................................................................................. 32 4.4.2 Login Screen ............................................................................................ 32 4.4.3 Registration Screen .................................................................................. 33 4.4.4 Forget Password Screen ........................................................................... 34 4.4.5 Home Screen ............................................................................................ 35 4.4.6 Create Activity Screen ............................................................................. 36 4.4.7Activities Screen ...................................................................................... 38 \n", + "attributes": [ + { + "attribute": "section", + "value": "Chapter 4 – Results and Discussion (TOC)" + }, + { + "attribute": "page_number", + "value": "8" + }, + { + "attribute": "source", + "value": "Document TOC excerpt" + }, + { + "attribute": "confidence", + "value": "High" + }, + { + "attribute": "sentiment", + "value": "Neutral" + } + ] + }, + { + "type": "comment", + "insight": "The page references specific technology platforms (Firebase Realtime Database and Firebase Cloud Storage), indicating that the implementation discusses cloud-based data storage and retrieval.", + "content": "4.3 Results and Discussion for centralized database architecture with storage scheme for multiple activity tracking and reporting system ................................ 29 4.3.1 Firebase Realtime database ..................................................................... 30 4.3.2Firebase Cloud Storage ............................................................................. 31", + "attributes": [ + { + "attribute": "section", + "value": "Firebase integration in Database section" + }, + { + "attribute": "page_number", + "value": "29-31" + }, + { + "attribute": "source", + "value": "Document TOC excerpt" + }, + { + "attribute": "confidence", + "value": "High" + }, + { + "attribute": "sentiment", + "value": "Neutral" + } + ] + }, + { + "type": "comment", + "insight": "The TOC lists mobile app screens as part of the Results, including Introduction, Login, Registration, Forget Password, Home, Create Activity, and Activities, indicating a comprehensive mobile UI flow.", + "content": "4.4 Results and Discussions for the mobile application for managing monitored and reported activites. .......................................................................................... 31 4.4.1 Introduction Screen .................................................................................. 32 4.4.2 Login Screen ............................................................................................ 32 4.4.3 Registration Screen .................................................................................. 33 4.4.4 Forget Password Screen ........................................................................... 34 4.4.5 Home Screen ............................................................................................ 35 4.4.6 Create Activity Screen ............................................................................. 36 4.4.7Activities Screen ...................................................................................... 38 ", + "attributes": [ + { + "attribute": "section", + "value": "Mobile App Screens in TOC (4.4)" + }, + { + "attribute": "page_number", + "value": "31-38" + }, + { + "attribute": "source", + "value": "Document TOC excerpt" + }, + { + "attribute": "confidence", + "value": "High" + }, + { + "attribute": "sentiment", + "value": "Neutral" + } + ] + }, + { + "type": "comment", + "insight": "The page contains minor typographical irregularities (e.g., 4.3.2Firebase Cloud Storage and 'activites' spelling), suggesting possible OCR or formatting issues on this page.", + "content": "4.3.2Firebase Cloud Storage ............................................................................. 31 4.4.7Activities Screen ...................................................................................... 38 ", + "attributes": [ + { + "attribute": "issue", + "value": "Formatting/typo irregularities (missing space after 4.3.2, misspelling 'activites')" + }, + { + "attribute": "page_number", + "value": "31-38" + }, + { + "attribute": "source", + "value": "Document TOC excerpt" + }, + { + "attribute": "confidence", + "value": "High" + }, + { + "attribute": "sentiment", + "value": "Neutral" + } + ] + }, + { + "type": "fact", + "insight": "Figure 3.1 shows a Raspberry Pi 3 Controller; page 15.", + "content": "Figure 3.1:Raspberry Pi 3 Controller 15", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 3.1" + }, + { + "attribute": "description", + "value": "Raspberry Pi 3 Controller" + }, + { + "attribute": "page_number", + "value": "15" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 3.2 shows the Raspberry Pi Camera Module V2; page 16.", + "content": "Figure. 3.2: Raspberry Pi Camera Module V2 16", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 3.2" + }, + { + "attribute": "description", + "value": "Raspberry Pi Camera Module V2" + }, + { + "attribute": "page_number", + "value": "16" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 3.3 shows an AI-based multi-function activity monitoring and reporting system; page 16.", + "content": "Figure 3.3: AI based multi-function activity monitoring and reporting system 16", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 3.3" + }, + { + "attribute": "description", + "value": "AI based multi-function activity monitoring and reporting system" + }, + { + "attribute": "page_number", + "value": "16" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 3.4 shows the circuit design for the multi-function activity monitoring and reporting system; page 18.", + "content": "Figure 3.4: Circuit design for multi-function activity monitoring and reporting system 18", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 3.4" + }, + { + "attribute": "description", + "value": "Circuit design for multi-function activity monitoring and reporting system" + }, + { + "attribute": "page_number", + "value": "18" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 3.5 shows the flowchart for video acquisition; page 19.", + "content": "Figure 3.5: Flowchart for video acquisition 19", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 3.5" + }, + { + "attribute": "description", + "value": "Flowchart for video acquisition" + }, + { + "attribute": "page_number", + "value": "19" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 3.6 shows the flow chart for detecting and recognizing activities; page 20.", + "content": "Figure 3.6: Flow chart for detecting and recognizing activities 20", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 3.6" + }, + { + "attribute": "description", + "value": "Flow chart for detecting and recognizing activities" + }, + { + "attribute": "page_number", + "value": "20" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 3.7 shows the Single Shot Detector (SSD) Framework; page 21.", + "content": "Figure 3.7: Single Shot Detector (SSD) Framework 21", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 3.7" + }, + { + "attribute": "description", + "value": "Single Shot Detector (SSD) Framework" + }, + { + "attribute": "page_number", + "value": "21" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 3.8 (first entry) shows the flowchart for the multi-function activity monitoring and reporting system; page 22.", + "content": "Figure 3.8: Flowchart for multi-function activity monitoring and reporting system 22", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 3.8" + }, + { + "attribute": "description", + "value": "Flowchart for multi-function activity monitoring and reporting system" + }, + { + "attribute": "page_number", + "value": "22" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 3.8 (second entry) shows the centralized database architecture with storage scheme for multiple activity tracking and reporting system; page 23.", + "content": "Figure 3.8: Centralized database architecture with storage scheme for multiple activity tracking and reporting system 23", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 3.8 (2)" + }, + { + "attribute": "description", + "value": "Centralized database architecture with storage scheme for multiple activity tracking and reporting system" + }, + { + "attribute": "page_number", + "value": "23" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 3.9 illustrates React Navigation; page 25.", + "content": "Figure 3.9: Illustration of React Navigation 25", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 3.9" + }, + { + "attribute": "description", + "value": "Illustration of React Navigation" + }, + { + "attribute": "page_number", + "value": "25" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 4.1 shows motion detection directly from testing the Pi Camera on the desk; page 27.", + "content": "Figure 4.1: Motion detection directly from testing the Pi Camera on desk 27", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 4.1" + }, + { + "attribute": "description", + "value": "Motion detection directly from testing the Pi Camera on desk" + }, + { + "attribute": "page_number", + "value": "27" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 4.2 shows motion detection from Raspberry Pi Camera placed at the top; page 28.", + "content": "Figure 4.2: Motion detection from Raspberry Pi Camera placed at the top 28", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 4.2" + }, + { + "attribute": "description", + "value": "Motion detection from Raspberry Pi Camera placed at the top" + }, + { + "attribute": "page_number", + "value": "28" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 4.3 shows Body Posture Detection of a person sitting down; page 29.", + "content": "Figure 4.3: Body Posture Detection of a person sitting down 29", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 4.3" + }, + { + "attribute": "description", + "value": "Body Posture Detection of a person sitting down" + }, + { + "attribute": "page_number", + "value": "29" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 4.5 shows the Activities collection table on Firebase; page 30.", + "content": "Figure 4.5: Activities collection table on Firebase 30", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 4.5" + }, + { + "attribute": "description", + "value": "Activities collection table on Firebase" + }, + { + "attribute": "page_number", + "value": "30" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 4.6 shows the Camera collection table on Firebase; page 30.", + "content": "Figure 4.6: Camera collection table on Firebase 30", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 4.6" + }, + { + "attribute": "description", + "value": "Camera collection table on Firebase" + }, + { + "attribute": "page_number", + "value": "30" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 4.7 shows the Tracked collection table on Firebase; page 31.", + "content": "Figure 4.7: Tracked collection table on Firebase 31", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 4.7" + }, + { + "attribute": "description", + "value": "Tracked collection table on Firebase" + }, + { + "attribute": "page_number", + "value": "31" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 4.7 (Firebase Cloud Storage) shows Firebase Cloud Storage; page 31.", + "content": "Figure 4.7: Firebase Cloud Storage 31", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 4.7 (2)" + }, + { + "attribute": "description", + "value": "Firebase Cloud Storage" + }, + { + "attribute": "page_number", + "value": "31" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 4.8 shows the Introduction Screen; page 32.", + "content": "Figure 4.8: Introduction Screen 32", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 4.8" + }, + { + "attribute": "description", + "value": "Introduction Screen" + }, + { + "attribute": "page_number", + "value": "32" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 4.9 shows the Login Screen; page 33.", + "content": "Figure 4.9: Login Screen 33", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 4.9" + }, + { + "attribute": "description", + "value": "Login Screen" + }, + { + "attribute": "page_number", + "value": "33" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 4.10 shows the Registration Screen; page 34.", + "content": "Figure 4.10: Registration Screen 34", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "attribute_description", + "value": "N/A" + }, + { + "attribute": "figure_label", + "value": "Figure 4.10" + }, + { + "attribute": "description", + "value": "Registration Screen" + }, + { + "attribute": "page_number", + "value": "34" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 4.11 shows the Forgot Password Screen; page 35.", + "content": "Figure 4.11: Forgot Password Screen 35", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 4.11" + }, + { + "attribute": "description", + "value": "Forgot Password Screen" + }, + { + "attribute": "page_number", + "value": "35" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Figure 4.12 shows the Home Screen; page 36.", + "content": "Figure 4.12: Home Screen 36", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 4.12" + }, + { + "attribute": "description", + "value": "Home Screen" + }, + { + "attribute": "page_number", + "value": "36" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "comment", + "insight": "There are two entries labeled Figure 3.8 on this page and two entries labeled Figure 4.7, indicating possible numbering duplication or mislabeling in the figure list.", + "content": "Figure 3.8: Flowchart for multi-function activity monitoring and reporting system 22", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 3.8" + }, + { + "attribute": "description", + "value": "Flowchart for multi-function activity monitoring and reporting system" + }, + { + "attribute": "page_number", + "value": "22" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "comment", + "insight": "The page includes a sequence of Figure numbers from 3.x to 4.x, indicating a document structure that covers both theory/design (3.x) and UI implementation (4.x).", + "content": "Figure 4.12: Home Screen 36", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "figure_label", + "value": "Figure 4.12" + }, + { + "attribute": "description", + "value": "Home Screen" + }, + { + "attribute": "page_number", + "value": "36" + }, + { + "attribute": "source_page", + "value": "9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Page 10 contains a LIST OF FIGURES listing figures 3.1 through 4.12 with captions and page numbers, indicating the document's structure uses chapters 3 and 4.", + "content": "LIST OF FIGURES Figure 3.1:Raspberry Pi 3 Controller 15 Figure. 3.2: Raspberry Pi Camera Module V2 16 Figure 3.3: AI based multi-function activity monitoring and reporting system 16 Figure 3.4: Circuit design for multi-function activity monitoring and reporting system 18 Figure 3.5: Flowchart for video acquisition 19 Figure 3.6: Flow chart for detecting and recognizing activities 20 Figure 3.7: Single Shot Detector (SSD) Framework 21 Figure 3.8: Flowchart for multi-function activity monitoring and reporting system 22 Figure 3.8: Centralized database architecture with storage scheme for multiple activity tracking and reporting system 23 Figure 3.9: Illustration of React Navigation 25 Figure 4.1: Motion detection directly from testing the Pi Camera on desk 27 Figure 4.2: Motion detection from Raspberry Pi Camera placed at the top 28 Figure 4.3: Body Posture Detection of a person sitting down 29 Figure 4.5: Activities collection table on Firebase 30 Figure 4.6: Camera collection table on Firebase 30 Figure 4.7: Tracked collection table on Firebase 31 Figure 4.7: Firebase Cloud Storage 31 Figure 4.8: Introduction Screen 32 Figure 4.9: Login Screen 33 Figure 4.10: Registration Screen 34 Figure 4.11: Forgot Password Screen 35 Figure 4.12: Home Screen 36", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "page", + "value": "10" + }, + { + "attribute": "range", + "value": "3.1-4.12" + }, + { + "attribute": "notes", + "value": "Two entries labelled Figure 3.8; 4.7 appears twice; minor typographical spacing differences present." + }, + { + "attribute": "confidence", + "value": "High" + } + ] + }, + { + "type": "fact", + "insight": "Figures 3.1-3.4 depict hardware and system design components: Raspberry Pi 3 Controller, Raspberry Pi Camera Module V2, AI-based multi-function activity monitoring system, and circuit design.", + "content": "Figure 3.1:Raspberry Pi 3 Controller 15 Figure. 3.2: Raspberry Pi Camera Module V2 16 Figure 3.3: AI based multi-function activity monitoring and reporting system 16 Figure 3.4: Circuit design for multi-function activity monitoring and reporting system 18", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "range", + "value": "3.1-3.4" + }, + { + "attribute": "confidence", + "value": "High" + } + ] + }, + { + "type": "fact", + "insight": "Figures 3.5-3.7 cover processing and software frameworks: flowchart for video acquisition, flowchart for detecting/ recognizing activities, and the Single Shot Detector (SSD) Framework.", + "content": "Figure 3.5: Flowchart for video acquisition 19 Figure 3.6: Flow chart for detecting and recognizing activities 20 Figure 3.7: Single Shot Detector (SSD) Framework 21", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "range", + "value": "3.5-3.7" + }, + { + "attribute": "confidence", + "value": "High" + } + ] + }, + { + "type": "fact", + "insight": "There appears to be multiple entries labeled Figure 3.8 and a separate Figure 3.9: 3.8 is listed twice with different captions (Flowchart for system; Centralized database architecture), and 3.9 is Illustration of React Navigation.", + "content": "Figure 3.8: Flowchart for multi-function activity monitoring and reporting system 22 Figure 3.8: Centralized database architecture with storage scheme for multiple activity tracking and reporting system 23 Figure 3.9: Illustration of React Navigation 25", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "range", + "value": "3.8-3.9" + }, + { + "attribute": "confidence", + "value": "High" + } + ] + }, + { + "type": "fact", + "insight": "Chapter 4 figures cover motion detection experiments, body posture detection, Firebase data tables, Firebase Cloud Storage, and various app UI screens.", + "content": "Figure 4.1: Motion detection directly from testing the Pi Camera on desk 27 Figure 4.2: Motion detection from Raspberry Pi Camera placed at the top 28 Figure 4.3: Body Posture Detection of a person sitting down 29 Figure 4.5: Activities collection table on Firebase 30 Figure 4.6: Camera collection table on Firebase 30 Figure 4.7: Tracked collection table on Firebase 31 Figure 4.7: Firebase Cloud Storage 31 Figure 4.8: Introduction Screen 32 Figure 4.9: Login Screen 33 Figure 4.10: Registration Screen 34 Figure 4.11: Forgot Password Screen 35 Figure 4.12: Home Screen 36", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "range", + "value": "4.1-4.12" + }, + { + "attribute": "confidence", + "value": "High" + } + ] + }, + { + "type": "comment", + "insight": "There are typographical inconsistencies/duplications in the list (e.g., two Figure 3.8 entries and two Figure 4.7 entries), and minor spacing irregularities in the captions.", + "content": "There appears to be numbering duplications (e.g., two entries labeled Figure 3.8 and two entries labeled Figure 4.7) and spacing differences in captions.", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "page", + "value": "10" + }, + { + "attribute": "confidence", + "value": "Medium" + } + ] + }, + { + "type": "fact", + "insight": "Overall, page 10 documents the List of Figures for a document describing a Raspberry Pi-based AI activity monitoring system with a Firebase backend and a React Native front end.", + "content": "LIST OF FIGURES ... Raspberry Pi 3 Controller ... React Navigation Introduction Screen ... Home Screen", + "attributes": [ + { + "attribute": "section", + "value": "List of Figures" + }, + { + "attribute": "page", + "value": "10" + }, + { + "attribute": "confidence", + "value": "High" + } + ] + }, + { + "type": "fact", + "insight": "AI-enabled multi-function activity monitoring and reporting system uses deep learning and computer vision to capture, analyze, predict activities, and report if instructed.", + "content": "The Artificial Intelligence-enabled multi-function activity monitoring and reporting system is an intelligent device that uses deep learning and computer vision to capture video feeds of a scene, analyse it, predict several activities and report if instructed to do so.", + "attributes": [ + { + "attribute": "section", + "value": "1.1 Background Study" + }, + { + "attribute": "source", + "value": "Chapter 1: Introduction" + } + ] + }, + { + "type": "fact", + "insight": "AI is a broad field that builds smart machines capable of performing tasks like humans or better.", + "content": "Artificial Intelligence (AI) is a wide branch that builds smart machines capable of performing tasks like humans or even better.", + "attributes": [ + { + "attribute": "section", + "value": "1.1 Background Study" + }, + { + "attribute": "source", + "value": "Chapter 1: Introduction" + } + ] + }, + { + "type": "fact", + "insight": "Deep learning mimics how the human brain works by using neural networks.", + "content": "Deep learning is a branch of AI that attempt to mimic how the human brain works by using neural networks.", + "attributes": [ + { + "attribute": "section", + "value": "1.1 Background Study" + }, + { + "attribute": "source", + "value": "Chapter 1: Introduction" + } + ] + }, + { + "type": "fact", + "insight": "Deep learning can work with images and text without requiring traditional structured data.", + "content": "This eliminates the need for using structured data like machine learning and can therefore work with images and text (Mishra & Gupta, 2017).", + "attributes": [ + { + "attribute": "section", + "value": "1.1 Background Study" + }, + { + "attribute": "source", + "value": "Chapter 1: Introduction" + } + ] + }, + { + "type": "fact", + "insight": "There is currently an inability to monitor and report several activities at once given security challenges, motivating the development of AI-enabled multi-activities systems.", + "content": "The inability to monitor and report several activities at a time considering the present world security challenges necessitates the development of the multi-faceted system.", + "attributes": [ + { + "attribute": "section", + "value": "1.2 Problem Statement" + }, + { + "attribute": "source", + "value": "Chapter 1: Introduction" + } + ] + }, + { + "type": "fact", + "insight": "Project objectives include design of a multi-function monitoring system, a centralized database architecture, and a mobile application.", + "content": "i. To design a multi-function activity monitoring and reporting system. ii. To design a centralized database architecture with storage scheme for multiple activity tracking and reporting system. iii. To design a mobile application for managing monitored and reported activities.", + "attributes": [ + { + "attribute": "section", + "value": "1.3 Aim and Objectives" + }, + { + "attribute": "source", + "value": "Chapter 1: Introduction" + } + ] + }, + { + "type": "fact", + "insight": "Scope and limitations include mobile prototype development and use of CNN and RNN with Faster RCNN; power and data constraints are noted.", + "content": "The scope of this project covers design, modelling and development of a mobile prototype of Artificial Intelligence-Enabled multi-functional monitoring and reporting System. Activity recognition is done by getting the video feed output from the camera mounted to the device, the output is then processed and passed to a video prediction algorithm to predict the activity at that point in time. The algorithm is based on deep learning which uses Convolutional Neural Networks (CNN) and Regression Neural Networks (RNN). The combination of these two creates a new algorithm known as Faster Regression Convolutional Neural Network (Faster RCNN) which is used to achieve this project. However, the envisaged limitations to this project is that of power consumption and large data collection.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations" + }, + { + "attribute": "source", + "value": "Chapter 1: Introduction" + } + ] + }, + { + "type": "fact", + "insight": "Model data size affects prediction quality in AI; limited data can lead to poor predictions.", + "content": "Any model fed with a limited amount of data, may not perform well and could lead to poor predictions.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations" + }, + { + "attribute": "source", + "value": "Chapter 1: Introduction" + } + ] + }, + { + "type": "fact", + "insight": "Justification highlights health and security applications, including monitoring patient status remotely and aiding in counter-insurgency monitoring.", + "content": "The potential sector which could benefit from this project includes Health and Security systems. Patient status can be monitored without the need of a medical personnel with this proposed system. Similarly, several insurgencies and bandit activities could be easily monitored and reported without physically intervention using this AI-enabled multi-functional activity monitoring and reporting system.", + "attributes": [ + { + "attribute": "section", + "value": "1.5 Justification" + }, + { + "attribute": "source", + "value": "Chapter 1: Introduction" + } + ] + }, + { + "type": "comment", + "insight": "Bandit activities are described as a major societal issue requiring AI-enabled self-reporting and monitoring.", + "content": "Bandit activities has been one of the major reported issues in society and requires an AI-enabled self-reporting and monitoring system.", + "attributes": [ + { + "attribute": "section", + "value": "1.2 Problem Statement" + }, + { + "attribute": "source", + "value": "Chapter 1: Introduction" + } + ] + }, + { + "type": "comment", + "insight": "The document outlines a five-chapter structure for the study (Introduction, literature review, methodology, data analysis, conclusions).", + "content": "Chapter One is structured into five chapters including Chapter One, Chapter Two, Chapter Three, Chapter Four and Chapter Five.", + "attributes": [ + { + "attribute": "section", + "value": "1.6 Structure of the Study" + }, + { + "attribute": "source", + "value": "Chapter 1: Introduction" + } + ] + }, + { + "type": "opinion", + "insight": "The author states that the goal is to make users feel more secure at homes, offices, and other places.", + "content": "not only to automatically monitor and report but also to make the users feel more secure about homes, offices and other places.", + "attributes": [ + { + "attribute": "section", + "value": "1.5 Justification" + }, + { + "attribute": "source", + "value": "Chapter 1: Introduction" + } + ] + }, + { + "type": "fact", + "insight": "The document cites several sources for background: Chaaraoui 2014, Wang 2019, Mishra & Gupta 2017, Tee 2015.", + "content": "Chaaraoui et al., 2014; Wang, 2019; Mishra & Gupta, 2017; Tee et al., 2015 are cited in the background and methodology discussions.", + "attributes": [ + { + "attribute": "section", + "value": "1.1 Background Study / 1.4 Scope and Limitations" + }, + { + "attribute": "source", + "value": "Chapter 1: Introduction" + } + ] + }, + { + "type": "fact", + "insight": "Chapter structure is stated to be five chapters, with topics including data presentation and analysis in Chapter Four.", + "content": "Chapter Four is data presentation and analysis. This Includes results and discussions.", + "attributes": [ + { + "attribute": "section", + "value": "1.6 Structure of the Study" + }, + { + "attribute": "source", + "value": "Chapter 1: Introduction" + } + ] + }, + { + "type": "fact", + "insight": "The project aims to design and implement an AI-based multi-function activity monitoring and reporting system with three objectives: (i) design a multi-function activity monitoring and reporting system, (ii) design a centralized database architecture with storage scheme for multiple activity tracking and reporting system, (iii) design a mobile application for managing monitored and reported activities.", + "content": "The aim of this project is to design and implement an Artificial Intelligent (AI) based multi-function Activity monitoring and reporting system with the following objectives: i. To design a multi-function activity monitoring and reporting system. ii. To design a centralized database architecture with storage scheme for multiple activity tracking and reporting system. iii. To design a mobile application for managing monitored and reported activites.", + "attributes": [ + { + "attribute": "section", + "value": "1.3 Aim and Objectives" + }, + { + "attribute": "source", + "value": "Page 12" + }, + { + "attribute": "reference", + "value": "—" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "The scope includes design, modelling and development of a mobile prototype of an AI-enabled multi-functional monitoring and reporting system, and it uses deep learning and computer vision to capture video feeds and predict activities.", + "content": "The scope of this project covers design, modelling and development of a mobile prototype of Artificial Intelligence-Enabled multi-functional monitoring and reporting System. This project focus is on multiple activities and events which include but not Limited to health, security and surveillance monitoring and reporting system using deep learning and computer vision to capture the video feeds and predict activities and report back to the user.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations of the Study" + }, + { + "attribute": "source", + "value": "Page 12" + } + ] + }, + { + "type": "fact", + "insight": "Activity recognition is performed by capturing the video feed from a camera mounted to the device, processing it, and passing it to a video prediction algorithm to predict the activity at that point in time.", + "content": "Activity recognition is done by getting the video feed output from the camera mounted to the device, the output is then processed and passed to a video prediction algorithm to predict the activity at that point in time.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations of the Study" + }, + { + "attribute": "source", + "value": "Page 12" + } + ] + }, + { + "type": "fact", + "insight": "The algorithm is based on deep learning and uses Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN); the combination creates Faster Regression Convolutional Neural Network (Faster RCNN)", + "content": "The algorithm is based on deep learning which uses Convolutional Neural Networks (CNN) and Regression Neural Networks (RNN). The combination of these two creates a new algorithm known as Faster Regression Convolutional Neural Network (Faster RCNN) which is used to achieve this project.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations of the Study" + }, + { + "attribute": "source", + "value": "Page 12" + } + ] + }, + { + "type": "fact", + "insight": "A limitation of the project is power consumption and large data collection requirements.", + "content": "However, the envisaged limitations to this project is that of power consumption and large data collection. The device makes use of a power supply to function effectively and the absence of power in the device makes the device inactive.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations of the Study" + }, + { + "attribute": "source", + "value": "Page 12" + } + ] + }, + { + "type": "fact", + "insight": "The amount of data collected and prepared for training determines how efficient the predictions will be; models trained with limited data may perform poorly.", + "content": "Moreso, in the field of AI, the amount of data collected, prepared and trained determines how efficient the predictions could be. Any model fed with a limited amount of data, may not perform well and could lead to poor predictions.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations of the Study" + }, + { + "attribute": "source", + "value": "Page 12" + } + ] + }, + { + "type": "comment", + "insight": "The text notes a major societal issue requiring an AI-enabled self-reporting and monitoring system.", + "content": "major reported issues in society and requires an AI-enabled self-reporting and monitoring system.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations of the Study" + }, + { + "attribute": "source", + "value": "Page 12" + } + ] + }, + { + "type": "fact", + "insight": "There is no known scenario in the literature where deep learning is used for activity monitoring with a self-reporting solution; a different approach has been used for monitoring elderly individuals using sensor and kinematic data (Tee et al., 2015).", + "content": "There’s no known scenario where deep learning is used for activity monitoring with a self-reporting solution but a different solution was found for monitoring the activities of elderly individuals using sensor and kinematic data ((Tee et al., 2015).", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations of the Study" + }, + { + "attribute": "source", + "value": "Page 12" + }, + { + "attribute": "reference", + "value": "Tee et al., 2015" + } + ] + }, + { + "type": "fact", + "insight": "The scope emphasizes that the project focuses on health, security, and surveillance monitoring and reporting.", + "content": "This project focus is on multiple activities and events which include but not Limited to health, security and surveillance monitoring and reporting system", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations of the Study" + }, + { + "attribute": "source", + "value": "Page 12" + } + ] + }, + { + "type": "fact", + "insight": "The project uses Faster RCNN as the core algorithm to achieve activity recognition.", + "content": "The combination of these two creates a new algorithm known as Faster Regression Convolutional Neural Network (Faster RCNN) which is used to achieve this project.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations of the Study" + }, + { + "attribute": "source", + "value": "Page 12" + } + ] + }, + { + "type": "fact", + "insight": "The project notes that the algorithm combines CNN and RNN in a Faster RCNN framework.", + "content": "The algorithm is based on deep learning which uses Convolutional Neural Networks (CNN) and Regression Neural Networks (RNN). The combination of these two creates a new algorithm known as Faster Regression Convolutional Neural Network (Faster RCNN) which is used to achieve this project.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations of the Study" + }, + { + "attribute": "source", + "value": "Page 12" + } + ] + }, + { + "type": "fact", + "insight": "The project acknowledges that large data collection and power supply are limitations for AI-enabled monitoring systems.", + "content": "However, the envisaged limitations to this project is that of power consumption and large data collection. The device makes use of a power supply to function effectively and the absence of power in the device makes the device inactive.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations of the Study" + }, + { + "attribute": "source", + "value": "Page 12" + } + ] + }, + { + "type": "opinion", + "insight": "There is an assertion that major societal issues require an AI-enabled self-reporting and monitoring system.", + "content": "major reported issues in society and requires an AI-enabled self-reporting and monitoring system.", + "attributes": [ + { + "attribute": "section", + "value": "Problem statement / motivation" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "There’s no known scenario where deep learning is used for activity monitoring with a self-reporting solution, but a different solution was found for monitoring the activities of elderly individuals using sensor and kinematic data (Tee et al., 2015).", + "content": "There’s no known scenario where deep learning is used for activity monitoring with a self-reporting solution but a different solution was found for monitoring the activities of elderly individuals using sensor and kinematic data ((Tee et al., 2015).", + "attributes": [ + { + "attribute": "section", + "value": "Intro / Background" + }, + { + "attribute": "source", + "value": "Tee et al., 2015" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "The aim of this project is to design and implement an Artificial Intelligent (AI) based multi-function Activity monitoring and reporting system.", + "content": "The aim of this project is to design and implement an Artificial Intelligent (AI) based multi-function Activity monitoring and reporting system.", + "attributes": [ + { + "attribute": "section", + "value": "1.3 Aim and Objectives" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "i. To design a multi-function activity monitoring and reporting system.", + "content": "i. To design a multi-function activity monitoring and reporting system.", + "attributes": [ + { + "attribute": "section", + "value": "1.3 Aim and Objectives" + }, + { + "attribute": "attribute", + "value": "objective" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "ii. To design a centralized database architecture with storage scheme for multiple activity tracking and reporting system.", + "content": "ii. To design a centralized database architecture with storage scheme for multiple activity tracking and reporting system.", + "attributes": [ + { + "attribute": "section", + "value": "1.3 Aim and Objectives" + }, + { + "attribute": "attribute", + "value": "objective" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "iii. To design a mobile application for managing monitored and reported activites.", + "content": "iii. To design a mobile application for managing monitored and reported activites.", + "attributes": [ + { + "attribute": "section", + "value": "1.3 Aim and Objectives" + }, + { + "attribute": "attribute", + "value": "objective" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "The scope of this project covers design, modelling and development of a mobile prototype of Artificial Intelligence-Enabled multi-functional monitoring and reporting System.", + "content": "The scope of this project covers design, modelling and development of a mobile prototype of Artificial Intelligence-Enabled multi-functional monitoring and reporting System.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "This project focus is on multiple activities and events which include but not Limited to health, security and surveillance monitoring and reporting system using deep learning and computer vision to capture the video feeds and predict activities and report back to the user.", + "content": "This project focus is on multiple activities and events which include but not Limited to health, security and surveillance monitoring and reporting system using deep learning and computer vision to capture the video feeds and predict activities and report back to the user.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "Activity recognition is done by getting the video feed output from the camera mounted to the device, the output is then processed and passed to a video prediction algorithm to predict the activity at that point in time.", + "content": "Activity recognition is done by getting the video feed output from the camera mounted to the device, the output is then processed and passed to a video prediction algorithm to predict the activity at that point in time.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "The algorithm is based on deep learning which uses Convolutional Neural Networks (CNN) and Regression Neural Networks (RNN).", + "content": "The algorithm is based on deep learning which uses Convolutional Neural Networks (CNN) and Regression Neural Networks (RNN).", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "The combination of these two creates a new algorithm known as Faster Regression Convolutional Neural Network (Faster RCNN) which is used to achieve this project.", + "content": "The combination of these two creates a new algorithm known as Faster Regression Convolutional Neural Network (Faster RCNN) which is used to achieve this project.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "However, the envisaged limitations to this project is that of power consumption and large data collection.", + "content": "However, the envisaged limitations to this project is that of power consumption and large data collection.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "The device makes use of a power supply to function effectively and the absence of power in the device makes the device inactive.", + "content": "The device makes use of a power supply to function effectively and the absence of power in the device makes the device inactive.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "Moreso, in the field of AI, the amount of data collected, prepared and trained determines how efficient the predictions could be.", + "content": "Moreso, in the field of AI, the amount of data collected, prepared and trained determines how efficient the predictions could be.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "Any model fed with a limited amount of data, may not perform well and could lead to poor predictions.", + "content": "Any model fed with a limited amount of data, may not perform well and could lead to poor predictions.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "comment", + "insight": "The text indicates the project includes health, security and surveillance monitoring, implying potential privacy considerations.", + "content": "This project focus is on multiple activities and events which include but not Limited to health, security and surveillance monitoring and reporting system using deep learning and computer vision to capture the video feeds and predict activities and report back to the user.", + "attributes": [ + { + "attribute": "section", + "value": "1.4 Scope and Limitations" + }, + { + "attribute": "page", + "value": "13" + }, + { + "attribute": "sentiment", + "value": "neutral" + } + ] + }, + { + "type": "fact", + "insight": "The project uses an AI-enabled multi-function activity recognition monitoring and reporting system to shift away from manual activity monitoring.", + "content": "In this project, manual activity monitoring and reporting system is given a phase shift with the use of an AI-enabled multi-function activity recognition monitoring and reporting system.", + "attributes": [ + { + "attribute": "section", + "value": "1.5 Justification" + }, + { + "attribute": "page", + "value": "14" + }, + { + "attribute": "source", + "value": "Page 14 of 56" + }, + { + "attribute": "confidence", + "value": "0.9" + }, + { + "attribute": "tone", + "value": "neutral" + } + ] + }, + { + "type": "fact", + "insight": "Potential sectors that could benefit include Health and Security systems.", + "content": "The potential sector which could benefit from this project includes Health and Security systems.", + "attributes": [ + { + "attribute": "section", + "value": "1.5 Justification" + }, + { + "attribute": "page", + "value": "14" + }, + { + "attribute": "source", + "value": "Page 14 of 56" + }, + { + "attribute": "confidence", + "value": "0.85" + } + ] + }, + { + "type": "fact", + "insight": "Patient status can be monitored without the need for medical personnel using the proposed system.", + "content": "Patient status can be monitored without the need of a medical personnel with this proposed system.", + "attributes": [ + { + "attribute": "section", + "value": "1.5 Justification" + }, + { + "attribute": "page", + "value": "14" + }, + { + "attribute": "source", + "value": "Page 14 of 56" + }, + { + "attribute": "confidence", + "value": "0.85" + } + ] + }, + { + "type": "fact", + "insight": "Several insurgencies and bandit activities could be monitored and reported without physical intervention using the AI-enabled system.", + "content": "Similarly, several insurgencies and bandit activities could be easily monitored and reported without physically intervention using this AI-enabled multi-functional activity monitoring and reporting system.", + "attributes": [ + { + "attribute": "section", + "value": "1.5 Justification" + }, + { + "attribute": "page", + "value": "14" + }, + { + "attribute": "source", + "value": "Page 14 of 56" + }, + { + "attribute": "confidence", + "value": "0.85" + } + ] + }, + { + "type": "opinion", + "insight": "Humans have an attention span shorter than a goldfish, making it impossible for an individual to monitor multiple activities at once.", + "content": "Since the human attention span is said to be shorter than a goldfish, it’s impossible for an individual of that nature to completely monitor and report several activities at a time.", + "attributes": [ + { + "attribute": "section", + "value": "1.5 Justification" + }, + { + "attribute": "page", + "value": "14" + }, + { + "attribute": "source", + "value": "Page 14 of 56" + }, + { + "attribute": "confidence", + "value": "0.80" + }, + { + "attribute": "sentiment", + "value": "critical of human capacity" + } + ] + }, + { + "type": "opinion", + "insight": "The project aims to automatically monitor/report and also to increase users' sense of security for homes, offices, and other places.", + "content": "In view of this, therefore, this project is designed not only to automatically monitor and report but also to make the users feel more secure about homes, offices and other places.", + "attributes": [ + { + "attribute": "section", + "value": "1.5 Justification" + }, + { + "attribute": "page", + "value": "14" + }, + { + "attribute": "source", + "value": "Page 14 of 56" + }, + { + "attribute": "confidence", + "value": "0.80" + }, + { + "attribute": "sentiment", + "value": "positive toward automation" + } + ] + }, + { + "type": "fact", + "insight": "The study is structured into five chapters: chapter one through chapter five.", + "content": "The project is structured into five chapters which include chapter one, chapter two, chapter three, chapter four and chapter five.", + "attributes": [ + { + "attribute": "section", + "value": "1.6 Structure of the Study" + }, + { + "attribute": "page", + "value": "14" + }, + { + "attribute": "source", + "value": "Page 14 of 56" + }, + { + "attribute": "confidence", + "value": "0.92" + } + ] + }, + { + "type": "fact", + "insight": "Chapter One is the Introductory Chapter and covers background study, problem statement, aim and objectives, scope and limitations, justification and structure.", + "content": "Chapter One is the Introductory Chapter and covers background study, problem statement, aim and objectives, scope and limitations, justification and structure.", + "attributes": [ + { + "attribute": "section", + "value": "1.6 Structure of the Study" + }, + { + "attribute": "page", + "value": "14" + }, + { + "attribute": "source", + "value": "Page 14 of 56" + }, + { + "attribute": "confidence", + "value": "0.92" + } + ] + }, + { + "type": "fact", + "insight": "Chapter Two is the theoretical literature review and shows analysis, synthesis, and evaluation of the project topic.", + "content": "Chapter Two is the theoretical literature review. This shows the analysis, synthesis and evaluation of the project topic.", + "attributes": [ + { + "attribute": "section", + "value": "1.6 Structure of the Study" + }, + { + "attribute": "page", + "value": "14" + }, + { + "attribute": "source", + "value": "Page 14 of 56" + }, + { + "attribute": "confidence", + "value": "0.90" + } + ] + }, + { + "type": "fact", + "insight": "Chapter Three is the research method and methodology, including methods, strategies, procedures, and materials used.", + "content": "Chapter Three is the research method and methodology. This includes different methods, strategies, procedures and materials used in this project.", + "attributes": [ + { + "attribute": "section", + "value": "1.6 Structure of the Study" + }, + { + "attribute": "page", + "value": "14" + }, + { + "attribute": "source", + "value": "Page 14 of 56" + }, + { + "attribute": "confidence", + "value": "0.90" + } + ] + }, + { + "type": "fact", + "insight": "Chapter Four is data presentation and analysis, including results and discussions.", + "content": "Chapter Four is the data presentation and analysis. This Includes results and discussions.", + "attributes": [ + { + "attribute": "section", + "value": "1.6 Structure of the Study" + }, + { + "attribute": "page", + "value": "14" + }, + { + "attribute": "source", + "value": "Page 14 of 56" + }, + { + "attribute": "confidence", + "value": "0.90" + } + ] + }, + { + "type": "fact", + "insight": "Chapter Five includes the conclusion and recommendation.", + "content": "Chapter Five includes the conclusion and recommendation.", + "attributes": [ + { + "attribute": "section", + "value": "1.6 Structure of the Study" + }, + { + "attribute": "page", + "value": "14" + }, + { + "attribute": "source", + "value": "Page 14 of 56" + }, + { + "attribute": "confidence", + "value": "0.90" + } + ] + }, + { + "type": "fact", + "insight": "Fast/Faster RCNN extracts region-independent features using convolutional layers initialized with discriminative pretraining for ImageNet classification, followed by a region-wise multilayer perceptron (MLP) for classification; they jointly optimize a softmax classifier and bounding-box regressors rather than training a softmax classifier, SVMs, and regressors in three separate stages; however, this approach is a memory hog because it uses MLP classifier topologies.", + "content": "The object detection pipeline has undergone significant evolution thanks to Fast RCNN and Faster RCNN (Deng, 2009; Girshick, 2015). Fast/Faster RCNN, a forerunner of the RCNN, extracts region-independent features using convolutional layers initialized with discriminative pretraining for ImageNet classification, followed by a region-wise multilayer perceptron (MLP) for classification (Ren, et al., 2015). Additionally, they jointly optimize a sofmax classifier and bounding-box repressors as opposed to training a sofmax classifier, SVMs, and repressors in three different stages. However, this approach is a memory hog because it uses MLP classifier topologies.", + "attributes": [ + { + "attribute": "section", + "value": "2.2 Fast-RCNN" + }, + { + "attribute": "references", + "value": "Deng 2009; Girshick 2015; Ren et al. 2015" + }, + { + "attribute": "notes", + "value": "Memory hog due to MLP classifier topologies" + } + ] + }, + { + "type": "fact", + "insight": "The core contributions in the Fast-RCNN/Faster RCNN discussion are three areas: (1) concatenation of the convolutional layer and the pooling layer to force the input size of the prevalent fully convolutional architectures to meet a set of requirements; (2) incorporating contemporary image classification networks such as ResNet and GoogleNet into Fast/Faster RCNN detection systems; (3) implementing skip connections similar to PVANET and FPN hybrids to combine intermediate outputs so that high-level semantic information and low-level visual features can be considered simultaneously.", + "content": "Our core work and contributions consist of the following three areas, which are based on a detailed analysis of the region wise feature classifier in Fast/Faster RCNN. First, we demonstrate that the concatenation of the convolutional layer and the pooling layer forces the input size of the prevalent fully convolutional architectures to meet a set of requirements. Second, based on a thorough examination of these fully convolutional architectures, we propose a method for incorporating contemporary, cutting-edge image classification networks, such as ResNet and various iterations of GoogleNet, into Fast/Faster RCNN detection systems (He, et al., 2016; Szegedy, et al., 2015). Finally, we implement the concept of skip connection similar to PVANET and the FPN hybrid that combines a number of intermediate outputs. As a result, both high-level semantic information and low-level visual features can be considered simultaneously (Kim, et al., 2016; Lin, et al., 2017).", + "attributes": [ + { + "attribute": "section", + "value": "2.2 Fast-RCNN" + }, + { + "attribute": "references", + "value": "He 2016; Szegedy 2015; Kim 2016; Lin 2017" + }, + { + "attribute": "notes", + "value": "Three areas: concatenation, integration of ResNet/GoogleNet, skip connections (PVANET/FPN)" + } + ] + }, + { + "type": "fact", + "insight": "PVANET and FPN-style skip connections enable combining intermediate outputs so that both high-level semantic information and low-level visual features can be considered simultaneously.", + "content": "Finally, we implement the concept of skip connection similar to PVANET and FPN hybrid that combines a number of intermediate outputs. As a result, both high-level semantic information and low-level visual features can be considered simultaneously (Kim, et al., 2016; Lin, et al., 2017).", + "attributes": [ + { + "attribute": "section", + "value": "2.2 Fast-RCNN" + }, + { + "attribute": "concept", + "value": "Skip connections PVANET/FPN" + } + ] + }, + { + "type": "fact", + "insight": "Computer vision, based on cameras, can detect, track, and classify objects and is a key technology for automatic monitoring and documentation without human intervention.", + "content": "Computer vision as the theory and technology of creating machines has the capability to detect, track, and classify objects. They are based on cameras as the main means of obtaining information about specific objects. A striking example of the application of this technology is the use of cameras for object detection, and for fixing violations. They allow monitoring of the observance of multi-functioning activities without human intervention, and capable of transmitting this information for due documentation on the configured network.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "notes", + "value": "Cameras as main info source; enables autonomous monitoring and reporting" + } + ] + }, + { + "type": "opinion", + "insight": "Publications in foreign literature indicate great prospects for using computer vision technology for multi-function activity monitoring.", + "content": "In addition, foreign publications on this topic also indicate great prospects for using this technology for the needs of multi-function activity monitoring.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "tone", + "value": "prospective/optimistic" + } + ] + }, + { + "type": "fact", + "insight": "The technology does not require regular presence of specialists and is well suited for monitoring complex infrastructure facilities that require constant monitoring, such as human activities, bridges, tunnels, and multi-level transport junctions.", + "content": "Since this technology does not require the regular presence of specialists in the field of monitoring at the facilities, its application is best suited for monitoring the functioning of complex infrastructure facilities that require constant monitoring of their condition, such as human activities, bridges and tunnels, multi-level transport junctions.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "use_case", + "value": "infrastructure monitoring" + }, + { + "attribute": "notes", + "value": "No regular specialist presence required" + } + ] + }, + { + "type": "fact", + "insight": "The technology enables automatic conversion of image or video data into objective information without direct human intervention.", + "content": "The goal of such systems is to reliably automatically convert image or video data into objective information without direct human intervention.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "goal", + "value": "automatic data conversion to objective information" + } + ] + }, + { + "type": "fact", + "insight": "There is ongoing work to create and modernize complexes based on stationary and mobile cameras.", + "content": "In addition, active work is underway to create and modernize already created complexes based on stationary and mobile cameras.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "projects", + "value": "stationary and mobile cameras" + } + ] + }, + { + "type": "comment", + "insight": "A sentence mentions algorithms that enable cameras to recognize and report various ongoing activities, but the sentence is cut off in this page.", + "content": "Another important area in the development of this technology is the creation of algorithms that allow cameras to recognize and report various kinds of ongoing activities which are complex for", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "note", + "value": "Content truncated on page" + } + ] + }, + { + "type": "fact", + "insight": "Demographic shifts due to longer life expectancy increase demand for personal autonomy support systems.", + "content": "Society is confronting a critical dilemma associated to greater life expectancy and a higher number of persons in dependent circumstances as a result of progress and demographic change.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "topic", + "value": "demographics / aging society" + }, + { + "attribute": "source", + "value": "Page 16" + } + ] + }, + { + "type": "fact", + "insight": "There is a significant demand for personal autonomy support systems.", + "content": "As a result, there is a significant demand for personal autonomy support systems.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "topic", + "value": "assistive autonomy technology" + }, + { + "attribute": "source", + "value": "Page 16" + } + ] + }, + { + "type": "fact", + "insight": "Chaaraoui et al. (2014) developed a vision-based home monitoring system enabling independent living for elderly and disabled individuals.", + "content": "Chaaraoui et al. (2014), developed a vision home system that allows elderly and disabled individuals to live independently at home while receiving care and safety services via vision-based monitoring.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "source", + "value": "Chaaraoui 2014" + } + ] + }, + { + "type": "fact", + "insight": "The vision-based system contributes to human behavior analysis and privacy protection.", + "content": "The system's specification is also offered, as well as innovative contributions to human behavior analysis and privacy protection.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "source", + "value": "Chaaraoui 2014" + } + ] + }, + { + "type": "fact", + "insight": "Experimental findings show excellent performance and support for multi-view and real-time execution.", + "content": "The behavior recognition method's experimental findings indicate excellent performance, as well as support for multi-view situations and real-time execution, both of which are necessary to provide the suggested services.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "source", + "value": "Chaaraoui 2014" + } + ] + }, + { + "type": "fact", + "insight": "The study discusses computer vision technologies, neural networks, and AI methodologies in monitoring.", + "content": "The study also discusses computer vision technologies, neural network techniques, and artificial intelligence methodologies in connection to the topic of monitoring.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "source", + "value": "Chaaraoui 2014" + } + ] + }, + { + "type": "fact", + "insight": "An intelligent monitoring system structure and a stationary monitoring complex based on video surveillance cameras were graphically represented.", + "content": "As a result, a graphical representation of the structure of an intelligent system for support and decision-making, as well as a block diagram of a stationary monitoring complex based on video surveillance cameras, was provided.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "source", + "value": "Chaaraoui 2014" + } + ] + }, + { + "type": "comment", + "insight": "The practicality and prospects of using such complexes for monitoring engineering infrastructure and their broader technological impact were discussed.", + "content": "The practicality and prospects of using such complexes for the monitoring needs of engineering infrastructure facilities, as well as the impact of the development of technologies used in them on global advancement in general, were discussed.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "source", + "value": "Chaaraoui 2014" + } + ] + }, + { + "type": "fact", + "insight": "Image processing is the process of converting images to digital form to enhance or extract information.", + "content": "Image processing is a way to convert an image to a digital aspect and perform certain functions on it, in order to get an enhanced image or extract other useful information from it.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "source", + "value": "Page 16" + } + ] + }, + { + "type": "fact", + "insight": "Image processing consists of three main steps: importing, analysis/management, and enhancement with feature detection.", + "content": "Image processing basically involves the following three steps, which are: Importing an image with an optical scanner or digital photography; Analysis and image management including data compression; and image enhancement and visual detection patterns such as satellite imagery.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "source", + "value": "Page 16" + } + ] + }, + { + "type": "fact", + "insight": "Python image-processing libraries commonly used include Scikit-image, OpenCV, Mahotas, SimpleITK, SciPy, Pillow, and Matplotlib.", + "content": "The following libraries are involved in performing Image processing in python: Scikit-image; OpenCV; Mahotas; SimplelTK; SciPy; Pillow; Matplotlib.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "source", + "value": "Page 16" + } + ] + }, + { + "type": "fact", + "insight": "OpenCV-Python is a widely used API with a C/C++ backend, facilitating robust computer vision programming.", + "content": "OpenCV (Open Source Computer Vision Library) among others is one of the most widely used libraries in computer programming. OpenCV-Python is an OpenCV Python API. OpenCV-Python is not only running, because the background has a code written in C/C++, but it is also easy to extract and distribute (due to Python folding in the front).", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "source", + "value": "Page 16-17" + } + ] + }, + { + "type": "fact", + "insight": "Home-monitoring systems using Raspberry Pi and Android smartphones can provide real-time in-house monitoring with motion detection and alerts.", + "content": "Every home owner should have a monitoring system in place for real-time monitoring of in-house activity. Surya and Ningsih (2019) developed a system that includes a Raspberry Pi as the computing core and an Android-based smartphone for monitoring.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "source", + "value": "Surya & Ningsih 2019" + } + ] + }, + { + "type": "fact", + "insight": "The Surya & Ningsih system uses a servo motor to move the camera and provides motion detection with alerting, achieving about 10.7 fps with two clients and under 1-second camera response.", + "content": "This system consists of a servo motor that can move the camera horizontally and vertically, a user-controllable video recording system, and a motion detection system that can alert users. System testing is done to obtain a real-time, dependable, and intelligent monitoring system by testing the network in terms of bandwidth characteristics and the number of clients in a network. According to test results with a maximum of two clients in one network, the system can take an average of 10.7 frames per second and has a camera movement response of less than one second.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Computer Vision" + }, + { + "attribute": "source", + "value": "Surya & Ningsih 2019" + } + ] + }, + { + "type": "fact", + "insight": "Data efficiency impacts performance in CV/AI benchmarks, per Mohana and Aradhya (2019).", + "content": "Mohana and Aradhya (2019) talked about how efficient data has impacted performance benchmarks in terms of speed and accuracy.", + "attributes": [ + { + "attribute": "section", + "value": "2.3-2.4 Computer Vision / AI" + }, + { + "attribute": "source", + "value": "Mohana & Aradhya 2019" + } + ] + }, + { + "type": "fact", + "insight": "CV and AI advances improve data visualization, processing, and analysis, enabling tasks like object detection and tracking for traffic vigilance.", + "content": "Computer vision (CV) and artificial intelligence (AI) have also improved data visualization, processing, and analysis. Major tasks such as object detection and tracking for traffic vigilance systems have been made possible by this technology.", + "attributes": [ + { + "attribute": "section", + "value": "2.4 Artificial Intelligence" + }, + { + "attribute": "source", + "value": "Page 16" + } + ] + }, + { + "type": "fact", + "insight": "CNN models for single object detection and YOLOv3 for multi-object detection are used on KITTI/COCO datasets.", + "content": "On the KITTI and COCO datasets, a convolutional neural network (CNN) model is constructed for single object detection on the urban vehicle dataset and YOLOv3 for multiple item detection on the YOLOv3 dataset.", + "attributes": [ + { + "attribute": "section", + "value": "2.4 Artificial Intelligence" + }, + { + "attribute": "source", + "value": "Page 16" + } + ] + }, + { + "type": "fact", + "insight": "Convolutional Neural Networks and deep learning enable rapid data processing and improved predictive accuracy in AI.", + "content": "Convolutional Neural Networks and deep learning artificial intelligence technologies are quickly evolving, primarily because AI processes large amounts of data much faster and makes predictions more accurately than humanly possible.", + "attributes": [ + { + "attribute": "section", + "value": "2.4 Artificial Intelligence" + }, + { + "attribute": "source", + "value": "Page 16" + } + ] + }, + { + "type": "fact", + "insight": "A noted disadvantage of AI is the cost of processing large data volumes (Indolia et al., 2018).", + "content": "As of this writing, the primary disadvantage of using AI was expensive to process the large amounts of data that AI programming requires (Indolia, et al., 2018).", + "attributes": [ + { + "attribute": "section", + "value": "2.4 Artificial Intelligence" + }, + { + "attribute": "source", + "value": "Indolia 2018" + } + ] + }, + { + "type": "fact", + "insight": "The digital world contains diverse data sources, including IoT, cybersecurity, mobile, business, social media, and health data.", + "content": "In the current age of the fourth industrial revolution, the digital world has a wealth of data, such as Internet of Things (IoT) data, cybersecurity data, mobile data, business data, social media data, and health data.", + "attributes": [ + { + "attribute": "section", + "value": "2.4 Artificial Intelligence" + }, + { + "attribute": "source", + "value": "Page 16" + } + ] + }, + { + "type": "fact", + "insight": "The document states society faces a dilemma due to increased life expectancy and more people in dependent circumstances as a result of progress and demographic change.", + "content": "Society is confronting a critical dilemma associated to greater life expectancy and a higher number of persons in dependent circumstances as a result of progress and demographic change.", + "attributes": [ + { + "attribute": "section", + "value": "Introduction / Societal context" + }, + { + "attribute": "source", + "value": "Document page 17 (page 16 in 0-indexed system)" + } + ] + }, + { + "type": "fact", + "insight": "There is a significant demand for personal autonomy support systems.", + "content": "There is a significant demand for personal autonomy support systems.", + "attributes": [ + { + "attribute": "section", + "value": "Introduction / Societal context" + }, + { + "attribute": "source", + "value": "Document page 17" + } + ] + }, + { + "type": "fact", + "insight": "Chaaraoui et al. (2014) developed a vision home system enabling elderly and disabled individuals to live independently at home with vision-based monitoring.", + "content": "Chaaraoui et al. (2014), developed a vision home system that allows elderly and disabled individuals to live independently at home while receiving care and safety services via vision-based monitoring.", + "attributes": [ + { + "attribute": "section", + "value": "Literature reference / Vision-based monitoring" + }, + { + "attribute": "source", + "value": "Chaaraoui et al., 2014" + } + ] + }, + { + "type": "fact", + "insight": "The study offers the system’s specifications and discusses contributions to human behavior analysis and privacy protection.", + "content": "The system's specification is also offered, as well as innovative contributions to human behavior analysis and privacy protection.", + "attributes": [ + { + "attribute": "section", + "value": "Literature reference / System specifications" + }, + { + "attribute": "source", + "value": "Chaaraoui et al., 2014" + } + ] + }, + { + "type": "fact", + "insight": "Experimental findings indicate excellent performance of the behavior recognition method, with support for multi-view situations and real-time execution.", + "content": "The behavior recognition method's experimental findings indicate excellent performance, as well as support for multi-view situations and real-time execution, both of which are necessary to provide the suggested services.", + "attributes": [ + { + "attribute": "section", + "value": "Literature results / Behavior recognition" + }, + { + "attribute": "source", + "value": "Chaaraoui et al., 2014" + } + ] + }, + { + "type": "fact", + "insight": "The study discusses computer vision technologies, neural networks, and AI methodologies in relation to monitoring.", + "content": "The study also discusses computer vision technologies, neural network techniques, and artificial intelligence methodologies in connection to the topic of monitoring.", + "attributes": [ + { + "attribute": "section", + "value": "Literature discussion / Technologies used" + }, + { + "attribute": "source", + "value": "Chaaraoui et al., 2014" + } + ] + }, + { + "type": "fact", + "insight": "Graphical representations were provided, including a structure diagram for an intelligent monitoring system and a block diagram for a stationary monitoring complex based on video surveillance cameras.", + "content": "As a result, a graphical representation of the structure of an intelligent system for support and decision-making, as well as a block diagram of a stationary monitoring complex based on video surveillance cameras, was provided.", + "attributes": [ + { + "attribute": "section", + "value": "Literature visuals / System diagrams" + }, + { + "attribute": "source", + "value": "Chaaraoui et al., 2014" + } + ] + }, + { + "type": "fact", + "insight": "The practicality and prospects of using such monitoring complexes for engineering infrastructure facilities, and the broader impact on global advancement, are discussed.", + "content": "The practicality and prospects of using such complexes for the monitoring needs of engineering infrastructure facilities, as well as the impact of the development of technologies used in them on global advancement in general, were discussed.", + "attributes": [ + { + "attribute": "section", + "value": "Literature discussion / Practicality & impact" + }, + { + "attribute": "source", + "value": "Chaaraoui et al., 2014" + } + ] + }, + { + "type": "fact", + "insight": "Image processing is defined as converting an image to a digital representation to perform functions for enhancement or information extraction.", + "content": "Image processing is a way to convert an image to a digital aspect and perform certain functions on it, in order to get an enhanced image or extract other useful information from it.", + "attributes": [ + { + "attribute": "section", + "value": "Image processing fundamentals" + }, + { + "attribute": "source", + "value": "Document page 17" + } + ] + }, + { + "type": "fact", + "insight": "Input to image processing is an image (or video frame) and output can be an image or features associated with the image.", + "content": "It is a type of signal time when the input is an image, such as a video frame or image and output can be an image or features associated with that image.", + "attributes": [ + { + "attribute": "section", + "value": "Image processing fundamentals" + }, + { + "attribute": "source", + "value": "Document page 17" + } + ] + }, + { + "type": "fact", + "insight": "Image processing is often described as treating images as two symbols and applying standard methods.", + "content": "Usually, the Image Processing system includes treating images as two equal symbols while using the set methods used.", + "attributes": [ + { + "attribute": "section", + "value": "Image processing fundamentals" + }, + { + "attribute": "source", + "value": "Document page 17" + } + ] + }, + { + "type": "fact", + "insight": "Image processing is a fast-growing technology with applications across various business sectors.", + "content": "It is one of the fastest growing technologies today, with its use in various business sectors.", + "attributes": [ + { + "attribute": "section", + "value": "Image processing market/industry trend" + }, + { + "attribute": "source", + "value": "Document page 17" + } + ] + }, + { + "type": "opinion", + "insight": "Graphic Design is described as the core of the research space within engineering and computer science industry.", + "content": "Graphic Design forms the core of the research space within the engineering and computer science industry as well.", + "attributes": [ + { + "attribute": "section", + "value": "Industry role / Graphic Design" + }, + { + "attribute": "source", + "value": "Document page 17" + } + ] + }, + { + "type": "fact", + "insight": "Image processing is defined as involving three main steps: Importing, Analysis and management (including data compression), and image enhancement and detection patterns (e.g., satellite imagery).", + "content": "Image processing basically involves the following three steps, which are: Importing an image with an optical scanner or digital photography; Analysis and image management including data compression; and image enhancement and visual detection patterns such as satellite imagery.", + "attributes": [ + { + "attribute": "section", + "value": "Image processing workflow" + }, + { + "attribute": "source", + "value": "Document page 17" + } + ] + }, + { + "type": "fact", + "insight": "The final output can be changed to an image or a report based on image analysis.", + "content": "It produces the final stage where the result can be changed to an image or report based on image analysis.", + "attributes": [ + { + "attribute": "section", + "value": "Image processing workflow" + }, + { + "attribute": "source", + "value": "Document page 17" + } + ] + }, + { + "type": "fact", + "insight": "Python image processing libraries include Scikit-image, OpenCV, Mahotas, SimpleLTK, SciPy, Pillow, and Matplotlib.", + "content": "The following libraries are involved in performing Image processing in python: Scikit-image; OpenCV; Mahotas; SimplelTK; SciPy; Pillow; Matplotlib.", + "attributes": [ + { + "attribute": "section", + "value": "Python libraries for image processing" + }, + { + "attribute": "source", + "value": "Document page 17" + } + ] + }, + { + "type": "fact", + "insight": "OpenCV is introduced as Open Source Computer Vision.", + "content": "OpenCV (Open Source Computer 17 ", + "attributes": [ + { + "attribute": "section", + "value": "OpenCV / Software library" + }, + { + "attribute": "source", + "value": "Document page 17" + } + ] + }, + { + "type": "fact", + "insight": "OpenCV-Python is an OpenCV Python API with a C/C++ backend, and the Python front end makes it easy to extract and distribute, contributing to more robust computer vision programs.", + "content": "OpenCV-Python is an OpenCV Python API. OpenCV-Python is not only running, because the background has a code written in C/C++, but it is also easy to extract and distribute (due to Python folding in the front). This makes it a good decision to make computer vision programs more robust.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Vision/ OpenCV and monitoring (OpenCV-Python description)" + }, + { + "attribute": "source", + "value": "OpenCV-Python description on page" + }, + { + "attribute": "date", + "value": "n/a" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "opinion", + "insight": "The text advocates that every home owner should have a monitoring system for real-time in-house activity monitoring.", + "content": "Every home owner should have a monitoring system in place for real-time monitoring of in-house activity.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Vision/ Applications (home monitoring advocacy)" + }, + { + "attribute": "source", + "value": "Text on page" + }, + { + "attribute": "date", + "value": "n/a" + }, + { + "attribute": "sentiment", + "value": "positive" + } + ] + }, + { + "type": "fact", + "insight": "Surya and Ningsih (2019) developed a Raspberry Pi–based monitoring system with an Android smartphone, including a servo-driven camera, video recording, and motion detection.", + "content": "Surya and Ningsih (2019) developed a system that includes a Raspberry Pi as the computing core and an Android-based smartphone for monitoring. This system consists of a servo motor that can move the camera horizontally and vertically, a user-controllable video recording system, and a motion detection system that can alert users.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Vision/ Monitoring System" + }, + { + "attribute": "source", + "value": "Surya & Ningsih (2019)" + }, + { + "attribute": "date", + "value": "2019" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Network testing reported that with up to two clients in a network, the system achieved an average of 10.7 frames per second and a camera movement response under one second.", + "content": "According to test results with a maximum of two clients in one network, the system can take an average of 10.7 frames per second and has a camera movement response of less than one second.", + "attributes": [ + { + "attribute": "section", + "value": "2.3 Vision/ Monitoring System (test results)" + }, + { + "attribute": "source", + "value": "Surya & Ningsih (2019)" + }, + { + "attribute": "date", + "value": "2019" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Mohana and Aradhya (2019) discussed how efficient data impacts performance benchmarks in terms of speed and accuracy.", + "content": "Mohana and Aradhya (2019) talked about how efficient data has impacted performance benchmarks in terms of speed and accuracy.", + "attributes": [ + { + "attribute": "section", + "value": "2.3/2.4 AI impact on performance (citation)" + }, + { + "attribute": "source", + "value": "Mohana & Aradhya (2019)" + }, + { + "attribute": "date", + "value": "2019" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Computer vision and AI have improved data visualization, processing, and analysis, enabling tasks such as object detection and tracking for traffic vigilance systems.", + "content": "Computer vision (CV) and artificial intelligence (AI) have also improved data visualization, processing, and analysis. Major tasks such as object detection and tracking for traffic vigilance systems have been made possible by this technology.", + "attributes": [ + { + "attribute": "section", + "value": "2.4 AI/ CV impact" + }, + { + "attribute": "source", + "value": "General statement on page" + }, + { + "attribute": "date", + "value": "n/a" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "On the KITTI and COCO datasets, a CNN model is constructed for single object detection on the urban vehicle dataset, and YOLOv3 is used for multi-object detection.", + "content": "On the KITTI and COCO datasets, a convolutional neural network (CNN) model is constructed for single object detection on the urban vehicle dataset and YOLOv3 for multiple item detection on the YOLOv3 dataset.", + "attributes": [ + { + "attribute": "section", + "value": "2.4 AI/ CNN and YOLO references" + }, + { + "attribute": "datasets", + "value": "KITTI, COCO" + }, + { + "attribute": "methods", + "value": "CNN (single object), YOLOv3 (multi-object)" + }, + { + "attribute": "confidence", + "value": "high" + }, + { + "attribute": "source", + "value": "Page text" + } + ] + }, + { + "type": "opinion", + "insight": "Artificial intelligence and deep learning technologies are rapidly evolving because they process large data faster and predict more accurately than humans.", + "content": "Artificial Intelligence Convolutional Neural Networks and deep learning artificial intelligence technologies are quickly evolving, primarily because AI processes large amounts of data much faster and makes predictions more accurately than humanly possible.", + "attributes": [ + { + "attribute": "section", + "value": "2.4 AI/ Deep learning evolution (opinion)" + }, + { + "attribute": "source", + "value": "Page text" + }, + { + "attribute": "date", + "value": "n/a" + }, + { + "attribute": "sentiment", + "value": "positive" + } + ] + }, + { + "type": "fact", + "insight": "AI uses machine learning to convert data into actionable information.", + "content": "While the huge volume of data being created on a daily basis would bury a human researcher, AI applications uses machine learning to take the data and quickly turn it into actionable information.", + "attributes": [ + { + "attribute": "section", + "value": "2.4 AI/ ML usage" + }, + { + "attribute": "source", + "value": "Page text" + }, + { + "attribute": "date", + "value": "n/a" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "The primary disadvantage of using AI, as cited here, is the cost of processing large amounts of data (Indolia, et al., 2018).", + "content": "As of this writing, the primary disadvantage of using AI was expensive to process the large amounts of data that AI programming requires (Indolia, et al., 2018).", + "attributes": [ + { + "attribute": "section", + "value": "2.4 AI/ limitations" + }, + { + "attribute": "source", + "value": "Indolia et al., 2018" + }, + { + "attribute": "date", + "value": "2018" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "The text notes a wealth of data in the digital world, including IoT, cybersecurity, mobile, business, social media, and health data.", + "content": "In the current age of the fourth industrial revolution, the digital world has a wealth of data, such as Internet of Things (IoT) data, cybersecurity data, mobile data, business data, social media data, and health data.", + "attributes": [ + { + "attribute": "section", + "value": "2.4 AI/ data types" + }, + { + "attribute": "source", + "value": "Page text" + }, + { + "attribute": "date", + "value": "n/a" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "AI/ML knowledge is key to developing smart, automated applications (Sarker 2021).", + "content": "analyze these data and develop the corresponding smart and automated applications, the knowledge of artificial intelligence (AI), particularly, machine learning (ML) is the key (Sarker, 2021).", + "attributes": [ + { + "attribute": "source", + "value": "Sarker 2021" + }, + { + "attribute": "section", + "value": "2.4.1 Machine Learning" + }, + { + "attribute": "date", + "value": "2021" + }, + { + "attribute": "confidence", + "value": "0.85" + } + ] + }, + { + "type": "fact", + "insight": "Deep learning is a part of ML and can analyze data at large scale with multiple application capabilities.", + "content": "Deep learning, as part of a broader family of machine learning methods, intelligently analyzes the data on a large scale with several application capabilities.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.1 Machine Learning" + }, + { + "attribute": "source", + "value": "General" + }, + { + "attribute": "confidence", + "value": "0.80" + } + ] + }, + { + "type": "fact", + "insight": "Nicosia (2019) developed an SVM and decision tree algorithm using a Udacity dataset to train for vehicle detection and tracking.", + "content": "Nicosia, (2019) developed a support vector machine and decision tree algorithm as dataset from Udacity were deployed with python programming language to create and train the algorithm for vehicle detection and tracking using machine learning.", + "attributes": [ + { + "attribute": "source", + "value": "Nicosia 2019" + }, + { + "attribute": "section", + "value": "2.4.1 Machine Learning" + }, + { + "attribute": "confidence", + "value": "0.80" + } + ] + }, + { + "type": "fact", + "insight": "The machine described is capable of monitoring and reporting multi-function activities.", + "content": "This machine is capable of monitoring and reporting multi-function activities.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.1 Machine Learning" + }, + { + "attribute": "source", + "value": "Nicosia 2019" + }, + { + "attribute": "confidence", + "value": "0.70" + } + ] + }, + { + "type": "fact", + "insight": "A conceptual model for the successful implementation of machine learning in organizations was presented in the article.", + "content": "A conceptual model for successful implementation of machine learning in organization was presented in this article.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.1 Machine Learning" + }, + { + "attribute": "source", + "value": "(article)" + }, + { + "attribute": "confidence", + "value": "0.75" + } + ] + }, + { + "type": "fact", + "insight": "Najafabadi et al. (2015) examined how deep learning is used to address problems in Big Data Analytics, including pattern extraction and fast information retrieval.", + "content": "Najafabadi, et al., (2015), investigated on how Deep Learning is being utilized for addressing some important problems in Big Data Analytics, including extracting complex patterns from massive volumes of data, semantic indexing, data tagging, fast information retrieval, and simplifying discriminative tasks.", + "attributes": [ + { + "attribute": "source", + "value": "Najafabadi 2015" + }, + { + "attribute": "section", + "value": "2.4.1 Machine Learning" + }, + { + "attribute": "confidence", + "value": "0.85" + } + ] + }, + { + "type": "fact", + "insight": "Big Data is increasingly important as organizations collect massive domain-specific information with potential relevance to national intelligence, cybersecurity, car detection, fraud detection, marketing, and medical informatics.", + "content": "Big Data has become important as many organizations both public and private have been collecting massive amounts of domain-specific information, which can contain useful information about problems such as national intelligence, cyber security, car detection, fraud detection, marketing, and medical informatics.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.1 Machine Learning" + }, + { + "attribute": "source", + "value": "Najafabadi 2015" + }, + { + "attribute": "confidence", + "value": "0.80" + } + ] + }, + { + "type": "fact", + "insight": "Deep Learning algorithms extract high-level abstractions as data representations via hierarchical learning, useful for Big Data Analytics when raw data are largely unlabeled.", + "content": "Deep Learning algorithms extract high-level, complex abstractions as data representations through a hierarchical learning process, which has made it a valuable tool for Big Data Analytics where raw data is largely unlabeled and uncategorized.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.1 Machine Learning" + }, + { + "attribute": "source", + "value": "Najafabadi 2015" + }, + { + "attribute": "confidence", + "value": "0.80" + } + ] + }, + { + "type": "fact", + "insight": "ML techniques are categorized into supervised, unsupervised, and reinforcement learning.", + "content": "Machine learning techniques are divided into three categories, which are:\ni. Supervised Machine Learning ...\nii. Unsupervised Machine Learning ...\niii. Reinforcement Learning ", + "attributes": [ + { + "attribute": "section", + "value": "2.4.1 Machine Learning" + }, + { + "attribute": "source", + "value": "Najafabadi 2015" + }, + { + "attribute": "confidence", + "value": "0.85" + } + ] + }, + { + "type": "fact", + "insight": "Supervised learning uses labeled data with predefined outputs; it includes algorithms like linear regression, logistic regression, multiclass classification, and SVM.", + "content": "The most widely used machine learning algorithms are supervised machine learning algorithms. A data scientist serves as a guide in this model, instructing the algorithm on what conclusions it should reach. In supervised learning, the algorithm is trained by a dataset that is already labeled and has a preset output, similar to how a youngster learns to identify fruits by remembering them in a picture book. Algorithms like linear and logistic regression, multiclass classification, and support vector machines are examples of supervised machine learning.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.1 Machine Learning" + }, + { + "attribute": "source", + "value": "Najafabadi 2015" + }, + { + "attribute": "confidence", + "value": "0.90" + } + ] + }, + { + "type": "fact", + "insight": "Unsupervised learning trains on unlabeled data without a clearly specified outcome.", + "content": "Unsupervised machine learning entails learning from data that lacks labels or a clearly specified outcome.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.1 Machine Learning" + }, + { + "attribute": "source", + "value": "Najafabadi 2015" + }, + { + "attribute": "confidence", + "value": "0.85" + } + ] + }, + { + "type": "fact", + "insight": "Reinforcement learning uses feedback after actions to guide learning; selection depends on data structure and use case.", + "content": "Data sets aren't labeled but, after performing an action or several actions, the AI system is given feedback. To choose an appropriate method for learning techniques requires the structure and volume of data, as well as the use case for further application which depends on either a supervised or unsupervised technique that was used.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.1 Machine Learning" + }, + { + "attribute": "source", + "value": "Najafabadi 2015" + }, + { + "attribute": "confidence", + "value": "0.85" + } + ] + }, + { + "type": "fact", + "insight": "Machine learning has been adopted across many industries, enabling goals like customer lifetime value, anomaly detection, dynamic pricing, predictive maintenance, image classification, and recommendations.", + "content": "enabling a wide range of corporate goals and use cases, including: customer lifetime value, anomaly detection, dynamic pricing, predictive maintenance, image classification, and recommendation engines.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.1 Machine Learning" + }, + { + "attribute": "source", + "value": "page 20 excerpt" + }, + { + "attribute": "confidence", + "value": "0.85" + } + ] + }, + { + "type": "fact", + "insight": "Deep Learning is a subset of ML consisting of a three-layer neural network with potential for added hidden layers to improve accuracy.", + "content": "Deep learning as a subset of machine learning is essential of a three-layer neural network. These neural networks aim to imitate the activity of the human brain by allowing it to learn from enormous amounts of data, albeit they fall far short of its capabilities. While single-layer neural network may produce approximate predictions, additional hidden layers could help to optimize and improve for accuracy.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.2 Deep Learning" + }, + { + "attribute": "source", + "value": "2.4.2 Deep Learning section" + }, + { + "attribute": "confidence", + "value": "0.85" + } + ] + }, + { + "type": "fact", + "insight": "Deep learning enables automation in many AI apps without human participation; examples include digital assistants and self-driving cars.", + "content": "Many artificial intelligence (AI) apps and services rely on deep learning to improve automation by executing analytical and physical activities without the need for human participation. Everyday products and services (such as digital assistants, voice-enabled TV remotes, such as self-driving cars, and credit card fraud detection) as well as upcoming innovations use deep learning technology.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.2 Deep Learning" + }, + { + "attribute": "source", + "value": "2.4.2 Deep Learning section" + }, + { + "attribute": "confidence", + "value": "0.80" + } + ] + }, + { + "type": "fact", + "insight": "Hordri et al. (2017) found deep learning suitable for better analysis and capable of handling large unlabeled data across fields.", + "content": "Hordri et al., (2017), reported that deep learning was suitable for better analysis, and it could learn enormous amounts of unlabeled data in various field.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.2 Deep Learning" + }, + { + "attribute": "source", + "value": "Hordri 2017" + }, + { + "attribute": "confidence", + "value": "0.80" + } + ] + }, + { + "type": "fact", + "insight": "Deep learning applications include speech recognition, image recognition, NLP, drug discovery, toxicology, CRM, recommendation systems, and bioinformatics.", + "content": "Deep learning finds its application in automatic speech recognition, image recognition, natural language processing, drug discovery and toxicology, customer relationship management, recommendation system and bioinformatics.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.2 Deep Learning" + }, + { + "attribute": "source", + "value": "Hordri 2017" + }, + { + "attribute": "confidence", + "value": "0.80" + } + ] + }, + { + "type": "fact", + "insight": "There are usability and adoptability concerns about deep learning systems (Kaluarachchi et al., 2021).", + "content": "Deep Learning has generated complexities in algorithms, and researchers and users have raised concerns regarding the usability and adoptability of Deep Learning systems (Kaluarachchi, et al., 2021).", + "attributes": [ + { + "attribute": "section", + "value": "2.4.2 Deep Learning" + }, + { + "attribute": "source", + "value": "Kaluarachchi 2021" + }, + { + "attribute": "confidence", + "value": "0.80" + } + ] + }, + { + "type": "opinion", + "insight": "These concerns, along with increasing human–AI interactions, have created the emerging field of Human-Centered Machine Learning (HCML).", + "content": "These concerns, coupled with the increasing human-artificial intelligence interactions, has created the emerging field that was Human-Centered Machine Learning (HCML).", + "attributes": [ + { + "attribute": "section", + "value": "2.4.2 Deep Learning" + }, + { + "attribute": "source", + "value": "Kaluarachchi 2021" + }, + { + "attribute": "confidence", + "value": "0.75" + } + ] + }, + { + "type": "fact", + "insight": "HCML involves collaborating with field domain experts to develop a working definition and identify research opportunities.", + "content": "Collaboration with field domain experts to develop a working definition for HCML was made. The topology of the HCML landscape by research gaps identification, were analyzed, highlighting conflicting interpretations, addressing current challenges, and presenting future HCML research opportunities.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.2 Deep Learning" + }, + { + "attribute": "source", + "value": "Kaluarachchi 2021" + }, + { + "attribute": "confidence", + "value": "0.75" + } + ] + }, + { + "type": "fact", + "insight": "Cloud-based architecture is defined as a system of network-connected distributed computing and storage servers providing virtual environments for various operations.", + "content": "A cloud based system comprises of collections of network-connected distributed computing and storage servers that provides virtual environments for different system operations, application", + "attributes": [ + { + "attribute": "section", + "value": "2.5 Cloud Based Architecture" + }, + { + "attribute": "source", + "value": "Page 19 excerpt" + }, + { + "attribute": "confidence", + "value": "0.80" + } + ] + }, + { + "type": "fact", + "insight": "Deep Learning enables a wide range of corporate goals and use cases, including customer lifetime value, anomaly detection, dynamic pricing, predictive maintenance, image classification, and recommendation engines.", + "content": "enabling a wide range of corporate goals and use cases, including: customer lifetime value, anomaly detection, dynamic pricing, predictive maintenance, image classification, and recommendation engines.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.1 Deep Learning Use Cases" + }, + { + "attribute": "source", + "value": "Page 20" + }, + { + "attribute": "date", + "value": "N/A" + }, + { + "attribute": "author", + "value": "N/A" + }, + { + "attribute": "topic", + "value": "Deep Learning Use Cases" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Deep learning is described as essential of a three-layer neural network, and adding hidden layers can optimize accuracy.", + "content": "Deep learning as a subset of machine learning is essential of a three-layer neural network. These neural networks aim to imitate the activity of the human brain by allowing it to learn from enormous amounts of data, albeit they fall far short of its capabilities. While single-layer neural network may produce approximate predictions, additional hidden layers could help to optimize and improve for accuracy.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.2 Deep Learning" + }, + { + "attribute": "source", + "value": "Page 20" + }, + { + "attribute": "date", + "value": "N/A" + }, + { + "attribute": "author", + "value": "Hordri et al., 2017" + }, + { + "attribute": "topic", + "value": "Deep Learning Theory" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "AI apps and services rely on deep learning to automate a wide range of activities without human participation.", + "content": "Many artificial intelligence (AI) apps and services rely on deep learning to improve automation by executing analytical and physical activities without the need for human participation.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.2 Deep Learning" + }, + { + "attribute": "source", + "value": "Page 20" + }, + { + "attribute": "date", + "value": "N/A" + }, + { + "attribute": "author", + "value": "Kaluarachchi et al., 2021" + }, + { + "attribute": "topic", + "value": "Automation via Deep Learning" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Examples of deep learning-enabled products include digital assistants, voice-enabled remotes, self-driving cars, and credit card fraud detection.", + "content": "Everyday products and services (such as digital assistants, voice-enabled TV remotes, such as self-driving cars, and credit card fraud detection) as well as upcoming innovations use deep learning technology.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.2 Deep Learning" + }, + { + "attribute": "source", + "value": "Page 20" + }, + { + "attribute": "date", + "value": "N/A" + }, + { + "attribute": "author", + "value": "N/A" + }, + { + "attribute": "topic", + "value": "Applications/Examples of DL" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "The deep learning algorithm uses gradient descent and backpropagation to optimize accuracy.", + "content": "The deep learning algorithm then changes and fits itself for accuracy via gradient descent and backpropagation,", + "attributes": [ + { + "attribute": "section", + "value": "2.4.2 Deep Learning" + }, + { + "attribute": "source", + "value": "Page 20" + }, + { + "attribute": "date", + "value": "N/A" + }, + { + "attribute": "author", + "value": "N/A" + }, + { + "attribute": "topic", + "value": "DL Algorithms" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Hordri et al. (2017) reported that deep learning is suitable for better analysis and can learn enormous amounts of unlabeled data in various fields.", + "content": "Hordri et al., (2017), reported that deep learning was suitable for better analysis, and it could learn enormous amounts unlabeled data in various field.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.2 Deep Learning" + }, + { + "attribute": "source", + "value": "Page 20" + }, + { + "attribute": "date", + "value": "2017" + }, + { + "attribute": "author", + "value": "Hordri et al." + }, + { + "attribute": "topic", + "value": "Deep Learning Applications/Analysis" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Deep Learning finds application in automatic speech recognition, image recognition, natural language processing, drug discovery and toxicology, CRM, recommender systems, and bioinformatics.", + "content": "Deep Learning find s application in automatic speech recognition, image recognition, natural language processing, drug discovery and toxicology, customer relationship management, recommendation system and bioinformatics.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.2 Deep Learning" + }, + { + "attribute": "source", + "value": "Page 20" + }, + { + "attribute": "date", + "value": "N/A" + }, + { + "attribute": "author", + "value": "N/A" + }, + { + "attribute": "topic", + "value": "DL Applications" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "There are concerns about the usability and adoptability of Deep Learning systems, as noted by researchers and users (Kaluarachchi, 2021).", + "content": "Deep Learning has generated complexities in algorithms, and researchers and users have raised concerns regarding the usability and adoptability of Deep Learning systems (Kaluarachchi, et al., 2021).", + "attributes": [ + { + "attribute": "section", + "value": "2.4.2 Deep Learning" + }, + { + "attribute": "source", + "value": "Page 20" + }, + { + "attribute": "date", + "value": "2021" + }, + { + "attribute": "author", + "value": "Kaluarachchi et al." + }, + { + "attribute": "topic", + "value": "HCML Concerns" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Human-Centered Machine Learning (HCML) is an emerging field that involves collaboration with domain experts to develop a working definition.", + "content": "These concerns, coupled with the increasing human-artificial intelligence (AI) interactions, has created the emerging field that was Human-Centered Machine Learning (HCML). Collaboration with field domain experts to develop a working definition for HCML was made.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.2 Deep Learning/HCML" + }, + { + "attribute": "source", + "value": "Page 20" + }, + { + "attribute": "date", + "value": "N/A" + }, + { + "attribute": "author", + "value": "N/A" + }, + { + "attribute": "topic", + "value": "HCML Definition" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "The HCML landscape topology has been analyzed to identify research gaps, highlight conflicting interpretations, address current challenges, and present future opportunities.", + "content": "The topology of the HCML landscape by research gaps identification, were analyzed, highlighting conflicting interpretations, addressing current challenges, and presenting future HCML research opportunities.", + "attributes": [ + { + "attribute": "section", + "value": "2.4.2 Deep Learning/HCML" + }, + { + "attribute": "source", + "value": "Page 20" + }, + { + "attribute": "date", + "value": "N/A" + }, + { + "attribute": "author", + "value": "N/A" + }, + { + "attribute": "topic", + "value": "HCML Landscape" + } + ] + }, + { + "type": "fact", + "insight": "Cloud-based Architecture is defined as a cloud-based system consisting of network-connected distributed computing and storage servers providing virtual environments for system operations and applications.", + "content": "2.5 Cloud Based Architecture A cloud based system comprises of collections of network-connected distributed computing and storage servers that provides virtual environments for different system operations, application", + "attributes": [ + { + "attribute": "section", + "value": "2.5 Cloud Based Architecture" + }, + { + "attribute": "source", + "value": "Page 20" + }, + { + "attribute": "date", + "value": "N/A" + }, + { + "attribute": "author", + "value": "N/A" + }, + { + "attribute": "topic", + "value": "Cloud Architecture" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Cloud-based systems provide virtual environments by pooling network-connected distributed computing and storage resources, enabling containers and computing services.", + "content": "A cloud based system comprises of collections of network-connected distributed computing and storage servers that provides virtual environments for different system operations, application containers, and computing services (Roehl, 2019).", + "attributes": [ + { + "attribute": "section", + "value": "Cloud Based Architecture" + }, + { + "attribute": "source", + "value": "Roehl 2019" + }, + { + "attribute": "page", + "value": "21" + }, + { + "attribute": "topic", + "value": "Cloud architecture basics" + } + ] + }, + { + "type": "fact", + "insight": "Using cloud platforms eliminates the need to acquire and maintain physical hardware.", + "content": "Software system functions on cloud platform by deploying it on the cloud platform's virtual machines, albeit the benefit is limited to eliminating the requirement for actual hardware acquisition and maintenance.", + "attributes": [ + { + "attribute": "section", + "value": "Cloud Based Architecture" + }, + { + "attribute": "source", + "value": "Roehl 2019" + }, + { + "attribute": "page", + "value": "21" + } + ] + }, + { + "type": "fact", + "insight": "Users must still configure computing environments (dependencies, networking, storage) on cloud servers.", + "content": "Within the cloud servers, users must still set up the appropriate computing environments, such as dependency libraries, networking, and storage system.", + "attributes": [ + { + "attribute": "section", + "value": "Cloud Based Architecture" + }, + { + "attribute": "source", + "value": "Roehl 2019" + }, + { + "attribute": "page", + "value": "21" + } + ] + }, + { + "type": "fact", + "insight": "Application containers help alleviate some setup by providing interfaces to computing environments for multiple deployments.", + "content": "Some of these responsibilities are alleviated by application containers, which give interfaces to computing environments that users do not have to set up separately for each deployment of the system software on a cloud platform.", + "attributes": [ + { + "attribute": "section", + "value": "Cloud Based Architecture" + }, + { + "attribute": "source", + "value": "Roehl 2019" + }, + { + "attribute": "page", + "value": "21" + } + ] + }, + { + "type": "fact", + "insight": "Cloud service models are categorized as IaaS, PaaS, FaaS, and SaaS.", + "content": "The use of cloud-based platforms can be classified into Infrastructure, Platform, Function, and software services:", + "attributes": [ + { + "attribute": "section", + "value": "Cloud Based Architecture" + }, + { + "attribute": "source", + "value": "Roehl 2019" + }, + { + "attribute": "page", + "value": "21" + } + ] + }, + { + "type": "fact", + "insight": "IaaS enables virtual machines for applications; example: Amazon EC2.", + "content": "i. Infrastructure as a Service (IaaS) This refers to the ability of a cloud platform to act as virtual machines for software applications that run on local servers. Amazon EC2 (Elastic Compute Cloud) is such an example, where users deploy software applications by launching instances of virtual servers, uploading the applications, and executing them.", + "attributes": [ + { + "attribute": "section", + "value": "Cloud Based Architecture" + }, + { + "attribute": "source", + "value": "Roehl 2019" + }, + { + "attribute": "page", + "value": "21" + } + ] + }, + { + "type": "fact", + "insight": "SaaS uses cloud platforms to provide software services (SaaS is introduced on this page; details continue on next page).", + "content": "iv. Software as a Service (SaaS) 22", + "attributes": [ + { + "attribute": "section", + "value": "Cloud Based Architecture" + }, + { + "attribute": "source", + "value": "Roehl 2019" + }, + { + "attribute": "page", + "value": "21" + } + ] + }, + { + "type": "fact", + "insight": "IaaS includes virtual machines that run on local servers and can use services like Amazon EC2; storage is provided via services such as Amazon S3.", + "content": "Amazon EC2 (Elastic Compute Cloud) is such an example, where users deploy software applications by launching instances of virtual servers, uploading the applications, and executing them. The virtual machines are distributed and are used as stable storage such for platforms such Amazon S3 (Simple Storage Service) buckets.", + "attributes": [ + { + "attribute": "section", + "value": "Cloud Based Architecture" + }, + { + "attribute": "source", + "value": "Roehl 2019" + }, + { + "attribute": "page", + "value": "21" + } + ] + }, + { + "type": "fact", + "insight": "PaaS manages computation environments for specialized applications; Google App Engine is an example.", + "content": "ii. Platform as a Service (PaaS) This manage the computation environment for specialized applications such as Web servlets. Google\\'s app engine is such an example, where users develop software programs using app engine development tools and deploy the program to the app engine for execution.", + "attributes": [ + { + "attribute": "section", + "value": "Cloud Based Architecture" + }, + { + "attribute": "source", + "value": "Roehl 2019" + }, + { + "attribute": "page", + "value": "21" + } + ] + }, + { + "type": "fact", + "insight": "In PaaS, users develop software similarly to IaaS but rely on the platform API for runtime support, networking, and storage.", + "content": "In this system, users need to develop complete software programs as in IaaS except that users rely on the API of the platform for runtime support, networking, and storage.", + "attributes": [ + { + "attribute": "section", + "value": "Cloud Based Architecture" + }, + { + "attribute": "source", + "value": "Roehl 2019" + }, + { + "attribute": "page", + "value": "21" + } + ] + }, + { + "type": "fact", + "insight": "FaaS enables stateless functions for high-throughput processing; examples include AWS Lambda and Azure Functions.", + "content": "iii. Function as a Service (FaaS) AWS Lambda and Azure Functions allow users to implement lightweight applications as stateless functions, which can be used for high throughput processing such as data transformation, altering, and event detection.", + "attributes": [ + { + "attribute": "section", + "value": "Cloud Based Architecture" + }, + { + "attribute": "source", + "value": "Roehl 2019" + }, + { + "attribute": "page", + "value": "21" + } + ] + }, + { + "type": "fact", + "insight": "FaaS functions connect to storage and network via APIs.", + "content": "FaaS functions connect to storage and network through APIs.", + "attributes": [ + { + "attribute": "section", + "value": "Cloud Based Architecture" + }, + { + "attribute": "source", + "value": "Roehl 2019" + }, + { + "attribute": "page", + "value": "21" + } + ] + }, + { + "type": "fact", + "insight": "FaaS is serverless with costs based on function calls.", + "content": "FaaS functions are considered server less since there are no dedicated servers allocated for running the functions and the cost is based on calls to the functions.", + "attributes": [ + { + "attribute": "section", + "value": "Cloud Based Architecture" + }, + { + "attribute": "source", + "value": "Roehl 2019" + }, + { + "attribute": "page", + "value": "21" + } + ] + }, + { + "type": "fact", + "insight": "IaaS is defined as providing virtual machines for software applications on a cloud platform, with Amazon EC2 given as an example.", + "content": "Infrastructure as a Service (IaaS) This refers to the ability of a cloud platform to act as virtual machines for software applications that run on local servers. Amazon EC2 (Elastic Compute Cloud) is such an example, where users deploy software applications by launching instances of virtual servers, uploading the applications, and executing them.", + "attributes": [ + { + "attribute": "section", + "value": "Cloud computing models (IaaS)" + }, + { + "attribute": "source", + "value": "Page 22" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "PaaS is described as managing the computation environment for specialized applications, with Google App Engine as an example.", + "content": "Platform as a Service (PaaS) This manage the computation environment for specialized applications such as Web servlets. Google\\'s app engine is such an example, where users develop software programs using app engine development tools and deploy the programs to the app engine for execution.", + "attributes": [ + { + "attribute": "section", + "value": "Cloud computing models (PaaS)" + }, + { + "attribute": "source", + "value": "Page 22" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "FaaS (Function as a Service) is exemplified by AWS Lambda and Azure Functions, enabling lightweight, stateless functions with pay-per-call cost.", + "content": "iii. Function as a Service (FaaS) AWS Lambda and Azure Functions allow users to implement lightweight applications as stateless functions, which can be used for high throughput processing such as data transformation, altering, and event detection. FaaS functions connect to storage and network through APIs. FaaS functions are considered server less since there are no dedicated servers allocated for running the functions and the cost is based on calls to the functions.", + "attributes": [ + { + "attribute": "section", + "value": "Cloud computing models (FaaS)" + }, + { + "attribute": "source", + "value": "Page 21-22" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "SaaS refers to software applications hosted by cloud servers such as Dropbox, Microsoft Office 365, and Google Apps that deliver functionalities over the network.", + "content": "This refers to software applications hosted by cloud servers such as Dropbox, Microsoft Oce 365, and Google Apps that deliver functionalities over the network.", + "attributes": [ + { + "attribute": "section", + "value": "Cloud computing models (SaaS)" + }, + { + "attribute": "source", + "value": "Page 22" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "The document states that the system uses Function Service for data transformation and event-based architecture consistent with IoT Hub and Stream Analytics services.", + "content": "For data transformation, modification, and event monitoring, this system employs Function Service, to serve as an event-based architecture. In view of its flexibility and scalability of resources, and is used for customizable IoT applications that is handled by the as IoT Hub and Stream Analytics services.", + "attributes": [ + { + "attribute": "section", + "value": "Cloud computing models (FaaS/SaaS) and IoT integration" + }, + { + "attribute": "source", + "value": "Page 22" + }, + { + "attribute": "confidence", + "value": "medium" + } + ] + }, + { + "type": "comment", + "insight": "2.6 Summary of the Reviewed Literatures appears to summarize tools and technologies for AI-enabled monitoring systems, including CNN, OCR, and web UI.", + "content": "2.6 Summary of the Reviewed Literatures In general, several tools were discovered for the development of an artificial intelligence-enabled multi-enabled activity monitoring and reporting system which includes, computer vision, convolutional neural network (CNN), web app as user interface for smart monitoring and reporting.", + "attributes": [ + { + "attribute": "section", + "value": "2.6 Summary of Reviewed Literature" + }, + { + "attribute": "source", + "value": "Page 22" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "CNN has capabilities for image processing, image recognition, and object detection and segmentation.", + "content": "Convolutional neural network has the capability for image processing, image recognition, object detection & segmentation.", + "attributes": [ + { + "attribute": "section", + "value": "CNN capabilities" + }, + { + "attribute": "source", + "value": "Page 22" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "CNN uses a hierarchical model that funnels information to a fully-connected layer, and plays a major role in the system.", + "content": "The CNN uses a hierarchical model that builds a network, similar to a funnel, and then outputs a fully-connected layer in which all neurons are connected to each other and the output is processed. Hence, CNN plays a major role in the develop of this system.", + "attributes": [ + { + "attribute": "section", + "value": "CNN architecture and role" + }, + { + "attribute": "source", + "value": "Page 22" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Computer vision, as a branch of AI, enables computers to extract information from digital images and videos and to act or suggest based on that data.", + "content": "As a branch of artificial intelligence (AI), computer vision allows computers and systems to extract useful information from digital photos, videos, and other visual inputs, and to conduct actions or make suggestions based on that data.", + "attributes": [ + { + "attribute": "section", + "value": "Computer vision overview" + }, + { + "attribute": "source", + "value": "Page 22" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Computer vision enables tracking, identification, and data logging with cameras and algorithms, reducing time.", + "content": "With cameras, data, and algorithms, computer vision trains this system to do these tracking, identification, and data logging duties in much less time.", + "attributes": [ + { + "attribute": "section", + "value": "Applications of computer vision" + }, + { + "attribute": "source", + "value": "Page 22" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Systems trained to inspect manufacturing items can evaluate thousands of products or processes per minute and detect faults.", + "content": "In addition, a system trained to inspect items or monitor a manufacturing asset can evaluate thousands of products or processes every minute, detecting faults or abnormalities that are otherwise undetectable.", + "attributes": [ + { + "attribute": "section", + "value": "Industrial inspection capabilities" + }, + { + "attribute": "source", + "value": "Page 22" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "OCR automates data extraction from scanned documents and images, transforming text to machine-readable format for data processing, with reported 98–99% accuracy and productivity gains.", + "content": "In this system OCR is used to automate data extraction from a scanned document or image file and then transform the text to a machine-readable format for data processing. This saves time and resources that would be required to manage unsearchable data otherwise. OCR eliminates human data entry, saves resources by processing more data faster and with fewer resources, lowers errors by providing a 98 to 99 percent accuracy range, and boosts productivity.", + "attributes": [ + { + "attribute": "section", + "value": "OCR and data extraction" + }, + { + "attribute": "source", + "value": "Page 22" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "IoT technology enhances GPS devices to transmit data remotely and connect to other systems and sensors to collect ongoing activity identification.", + "content": "IoT technology enhances GPS devices to transmit data remotely and connect to other systems and sensors, for the sole purpose of collecting and transmitting comprehensive ongoing activity identification.", + "attributes": [ + { + "attribute": "section", + "value": "IoT integration with GPS and sensors" + }, + { + "attribute": "source", + "value": "Page 22" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "comment", + "insight": "The sentence fragment 'Although individuals can engage with the devices to set them up' appears incomplete on this page.", + "content": "Although individuals can engage with the devices to set them up,", + "attributes": [ + { + "attribute": "section", + "value": "Page 22 continuity notes" + }, + { + "attribute": "source", + "value": "Page 22" + }, + { + "attribute": "confidence", + "value": "medium" + } + ] + }, + { + "type": "fact", + "insight": "Gadgets perform the majority of the work without human participation, whether instructed or used to retrieve data.", + "content": "give them instructions, or retrieve data, the gadgets do the majority of the work without human participation.", + "attributes": [ + { + "attribute": "source", + "value": "Document page 23" + }, + { + "attribute": "page_number", + "value": "23" + }, + { + "attribute": "topic", + "value": "automation/gadgets" + }, + { + "attribute": "certainty", + "value": "high" + }, + { + "attribute": "tone", + "value": "neutral" + } + ] + }, + { + "type": "comment", + "insight": "Gadgets perform the majority of the work without human participation.", + "content": "give them instructions, or retrieve data, the gadgets do the majority of the work without human participation.", + "attributes": [ + { + "attribute": "page_number", + "value": "24" + }, + { + "attribute": "context", + "value": "automation description" + }, + { + "attribute": "tone", + "value": "neutral" + }, + { + "attribute": "source", + "value": "document page 24" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "comment", + "insight": "Humans interact with gadgets primarily by giving instructions or retrieving data, while gadgets execute tasks autonomously.", + "content": "give them instructions, or retrieve data, the gadgets do the majority of the work without human participation.", + "attributes": [ + { + "attribute": "page_number", + "value": "24" + }, + { + "attribute": "context", + "value": "human-automation workflow" + }, + { + "attribute": "tone", + "value": "neutral" + }, + { + "attribute": "source", + "value": "document page 24" + }, + { + "attribute": "confidence", + "value": "medium" + } + ] + }, + { + "type": "fact", + "insight": "AI-based activity detection/monitoring is applied across security, surveillance, health, and other sectors.", + "content": "Artificial intelligence in activity detection and monitoring systems has application in security, surveillance, health, and a variety of sectors, where activities are monitored leveraging RCNN, pose estimation, and internet of things (IoT) processes as its approach for developing an AI-based multi-function activity monitoring and reporting system with a mobile application to serve as a gateway for interfacing with the system.", + "attributes": [ + { + "attribute": "section", + "value": "3.1 Overview" + }, + { + "attribute": "chapter", + "value": "CHAPTER THREE" + }, + { + "attribute": "page", + "value": "25" + }, + { + "attribute": "topic", + "value": "AI-based activity monitoring applications" + } + ] + }, + { + "type": "fact", + "insight": "The approach uses RCNN, pose estimation, and IoT processes to monitor activities.", + "content": "where activities are monitored leveraging RCNN, pose estimation, and internet of things (IoT) processes as its approach for developing an AI-based multi-function activity monitoring and reporting system with a mobile application to serve as a gateway for interfacing with the system.", + "attributes": [ + { + "attribute": "section", + "value": "3.1 Overview" + }, + { + "attribute": "method", + "value": "RCNN, pose estimation, IoT" + } + ] + }, + { + "type": "fact", + "insight": "The system includes a mobile application that serves as a gateway to interface with the system.", + "content": "... AI-based multi-function activity monitoring and reporting system with a mobile application to serve as a gateway for interfacing with the system.", + "attributes": [ + { + "attribute": "section", + "value": "3.1 Overview" + }, + { + "attribute": "feature", + "value": "mobile gateway" + } + ] + }, + { + "type": "fact", + "insight": "OpenCV library provides real-time capabilities and extensive computer vision features relevant to the system.", + "content": "OpenCV (Open-Source Computer Vision Library) has rich plugins for computer vision, machine learning, and image processing, and it presently plays a key role in real-time operation, which is crucial to this AI based multi-function activity monitoring and reporting system.", + "attributes": [ + { + "attribute": "section", + "value": "3.2.1 Software Materials" + }, + { + "attribute": "tool", + "value": "OpenCV" + } + ] + }, + { + "type": "fact", + "insight": "OpenCV can recognize items and people, identify objects, classify actions, and perform various tracking and 3D-related tasks.", + "content": "It recognizes items, people, identify objects, classify human actions, track camera movements, track moving objects, extract 3D models of objects, produce 3D point clouds from stereo cameras, stitch images together to produce a high resolution image of an entire scene, facial movements, detect scenery, and establish markers to overlay it with a 3D model.", + "attributes": [ + { + "attribute": "section", + "value": "3.2.1 Software Materials" + }, + { + "attribute": "tool", + "value": "OpenCV" + } + ] + }, + { + "type": "fact", + "insight": "TensorFlow is adopted for its symbolic math toolkit with dataflow and differentiable programming for training and inference in deep neural networks; it is used for image classification in this work.", + "content": "TensorFlow is adopted in this project for use because it has a symbolic math toolkit that employs dataflow and differentiable programming to handle a variety of tasks related to deep neural network training and inference. It’s used for image classification in this research work.", + "attributes": [ + { + "attribute": "section", + "value": "3.2.1 Software Materials" + }, + { + "attribute": "tool", + "value": "TensorFlow" + } + ] + }, + { + "type": "fact", + "insight": "Raspberry Pi 3 is described as a low-cost, credit-card-sized computer that connects to a display and operates with a keyboard and mouse, suitable for learning Scratch and Python.", + "content": "The Raspberry Pi 3 has a low-cost, credit-card-sized computer that connects to a computer display or TV and operates with a regular keyboard and mouse. It is a powerful tiny computer that allows individuals of all ages to experiment with computing and learn to write in languages such as Scratch and Python.", + "attributes": [ + { + "attribute": "section", + "value": "3.2.2 Hardware Materials" + }, + { + "attribute": "subsection", + "value": "i. Raspberry Pi 3" + }, + { + "attribute": "page", + "value": "26" + }, + { + "attribute": "confidence", + "value": "0.82" + } + ] + }, + { + "type": "fact", + "insight": "Raspberry Pi 3 includes GPIO pins for general-purpose input/output, enabling physical computing and IoT exploration.", + "content": "It also has GPIO (general purpose input/output) pins for controlling electrical components for physical computing and exploring the Internet of Things (IoT) as in fig. 3.1.", + "attributes": [ + { + "attribute": "section", + "value": "3.2.2 Hardware Materials" + }, + { + "attribute": "subsection", + "value": "i. Raspberry Pi 3" + }, + { + "attribute": "feature", + "value": "GPIO pins" + }, + { + "attribute": "page", + "value": "26" + }, + { + "attribute": "confidence", + "value": "0.79" + } + ] + }, + { + "type": "comment", + "insight": "Figure 3.1 is referenced as the Raspberry Pi 3 Controller in the text.", + "content": "Figure 3.1:Raspberry Pi 3 Controller", + "attributes": [ + { + "attribute": "section", + "value": "3.2.2 Hardware Materials" + }, + { + "attribute": "subsection", + "value": "i. Raspberry Pi 3" + }, + { + "attribute": "page", + "value": "26" + }, + { + "attribute": "confidence", + "value": "0.75" + } + ] + }, + { + "type": "fact", + "insight": "Raspberry Pi Camera Module V2 is described as an 8MP camera module that connects directly to the Raspberry Pi to capture video feed.", + "content": "ii. Raspberry Pi Camera Module V2 This 8MP camera module connects straight to the Raspberry Pi to collect video feed.", + "attributes": [ + { + "attribute": "section", + "value": "3.2.2 Hardware Materials" + }, + { + "attribute": "subsection", + "value": "ii. Raspberry Pi Camera Module V2" + }, + { + "attribute": "page", + "value": "26" + }, + { + "attribute": "confidence", + "value": "0.78" + } + ] + }, + { + "type": "fact", + "insight": "The camera module is stated to support pixel static images and video.", + "content": "It has supports of pixel static images, , and video.", + "attributes": [ + { + "attribute": "section", + "value": "3.2.2 Hardware Materials" + }, + { + "attribute": "subsection", + "value": "ii. Raspberry Pi Camera Module V2" + }, + { + "attribute": "page", + "value": "26" + }, + { + "attribute": "confidence", + "value": "0.65" + } + ] + }, + { + "type": "fact", + "insight": "The document provides resolution-related specifications for the camera module (e.g., 3280 × 2464, 1080p, etc.), but the text is garbled and partially unreadable.", + "content": "It is 3280 × 2464 1080 𝑝 30720 𝑝 60 640 ×480 𝑝 90adopted for this design because this is the plug-and-play version of the Raspbian operating system, making it ideal for video recording, motion detection, and security applications.", + "attributes": [ + { + "attribute": "section", + "value": "3.2.2 Hardware Materials" + }, + { + "attribute": "subsection", + "value": "ii. Raspberry Pi Camera Module V2" + }, + { + "attribute": "page", + "value": "26" + }, + { + "attribute": "confidence", + "value": "0.42" + } + ] + }, + { + "type": "fact", + "insight": "The camera module is described as plug-and-play with the Raspbian operating system and suitable for video recording, motion detection, and security applications.", + "content": "adopted for this design because this is the plug-and-play version of the Raspbian operating system, making it ideal for video recording, motion detection, and security applications.", + "attributes": [ + { + "attribute": "section", + "value": "3.2.2 Hardware Materials" + }, + { + "attribute": "subsection", + "value": "ii. Raspberry Pi Camera Module V2" + }, + { + "attribute": "page", + "value": "26" + }, + { + "attribute": "confidence", + "value": "0.38" + } + ] + }, + { + "type": "fact", + "insight": "The Raspberry Pi Camera Module V2 is connected to the Raspberry Pi via a ribbon cable to the CSI (Camera Serial Interface) connector.", + "content": "Connect the provided ribbon cable to your Raspberry Pi's CSI (Camera Serial Interface) connector, and you're ready to go!", + "attributes": [ + { + "attribute": "section", + "value": "3.2.2 Hardware Materials" + }, + { + "attribute": "subsection", + "value": "ii. Raspberry Pi Camera Module V2" + }, + { + "attribute": "page", + "value": "26" + }, + { + "attribute": "confidence", + "value": "0.77" + } + ] + }, + { + "type": "fact", + "insight": "The camera module is described as small, with dimensions around 25 mm × 23 mm × 9 mm and a weight of about 3 g.", + "content": "The board itself is small, measuring around and 25 𝑚𝑚 × 23 𝑚𝑚 × 9 𝑚𝑚weighing little over 3g, making it ideal for mobile or other applications where size and weight", + "attributes": [ + { + "attribute": "section", + "value": "3.2.2 Hardware Materials" + }, + { + "attribute": "subsection", + "value": "ii. Raspberry Pi Camera Module V2" + }, + { + "attribute": "page", + "value": "26" + }, + { + "attribute": "confidence", + "value": "0.68" + } + ] + }, + { + "type": "fact", + "insight": "The page references Fig. 3.1: Raspberry Pi 3 Controller.", + "content": "Figure 3.1:Raspberry Pi 3 Controller", + "attributes": [ + { + "attribute": "section", + "value": "3.2.2 Hardware Materials" + }, + { + "attribute": "subsection", + "value": "i. Raspberry Pi 3" + }, + { + "attribute": "page", + "value": "26" + }, + { + "attribute": "confidence", + "value": "0.70" + } + ] + }, + { + "type": "comment", + "insight": "The page contains garbled text and missing or unclear numerical data related to camera resolutions and dimensions.", + "content": "It is 3280 × 2464 1080 𝑝 30720 𝑝 60 640 ×480 𝑝 90adopted for this design because this is the plug-and-play version of the Raspbian operating system, making it ideal for video recording, motion detection, and security applications.", + "attributes": [ + { + "attribute": "section", + "value": "3.2.2 Hardware Materials" + }, + { + "attribute": "subsection", + "value": "ii. Raspberry Pi Camera Module V2" + }, + { + "attribute": "page", + "value": "26" + }, + { + "attribute": "confidence", + "value": "0.50" + } + ] + }, + { + "type": "comment", + "insight": "The page includes formatting issues and incomplete data (e.g., missing measurement values) that affect data readability.", + "content": "The board itself is small, measuring around and 25 𝑚𝑚 × 23 𝑚𝑚 × 9 𝑚𝑚weighing little over 3g, making it ideal for mobile or other applications where size and weight", + "attributes": [ + { + "attribute": "section", + "value": "3.2.2 Hardware Materials" + }, + { + "attribute": "subsection", + "value": "ii. Raspberry Pi Camera Module V2" + }, + { + "attribute": "page", + "value": "26" + }, + { + "attribute": "confidence", + "value": "0.48" + } + ] + }, + { + "type": "fact", + "insight": "The sensor features an 8-megapixel native resolution and a fixed-focus lens (as referenced by fig.3.2).", + "content": "are critical. The sensor features an 8-megapixel native resolution and a fixed-focus lens as in fig.3.2.", + "attributes": [ + { + "attribute": "source", + "value": "Document page 27; Figure 3.2; Hardware description" + }, + { + "attribute": "section", + "value": "Camera hardware (Raspberry Pi Camera Module V2)" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Raspberry Pi Camera Module V2 is used as the image source.", + "content": "The sensor features an 8-megapixel native resolution and a fixed-focus lens as in fig.3.2. Figure. 3.2: Raspberry Pi Camera Module V2", + "attributes": [ + { + "attribute": "source", + "value": "Document page 27; Figure 3.2" + }, + { + "attribute": "section", + "value": "Camera hardware (Raspberry Pi Camera Module V2)" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "The system is described as an AI-based multi-function activity monitoring and reporting system (illustrated by Figure 3.3).", + "content": "Figure. 3.2: Raspberry Pi Camera Module V2 3.3 Method The block diagram in Figure 3.3 shows AI based multi-function activity monitoring and reporting system.", + "attributes": [ + { + "attribute": "source", + "value": "Document page 27; Figure 3.3" + }, + { + "attribute": "section", + "value": "AI-based system (Figure 3.3)" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Video processing is described as using image processing techniques to process a video feed, starting with image acquisition from a camera source and converting images to grayscale for processing.", + "content": "i. Video Processing This entails using image processing techniques to process a video feed. This takes a succession of photographs in a frame, starting with image acquisition, which entails collecting the images from a source, in our case the raspberry pi camera module v2. The next step is to convert the image to grayscale so that it can be processed.", + "attributes": [ + { + "attribute": "source", + "value": "Document page 27" + }, + { + "attribute": "section", + "value": "Video Processing" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "The model used for motion detection is the Single-Shot Multibox Detector (SSD) Mobilenet network, loaded via Keras/OpenCV.", + "content": "ii. Load Model Single-Shot multibox Detection (SSD) mobilenet network was utilized for this project. This was used to perform motion detection. The process involves loading the model using Keras/OpenCV to make it accessible for use.", + "attributes": [ + { + "attribute": "source", + "value": "Document page 27" + }, + { + "attribute": "section", + "value": "Model (SSD Mobilenet) for motion detection" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "comment", + "insight": "The page describes a hardware-software pipeline combining Raspberry Pi camera input with an SSD Mobilenet-based motion detection model for an AI monitoring system.", + "content": "3.3 Method The block diagram in Figure 3.3 shows AI based multi-function activity monitoring and reporting system.", + "attributes": [ + { + "attribute": "source", + "value": "Document page 27" + }, + { + "attribute": "section", + "value": "Overall system description" + }, + { + "attribute": "confidence", + "value": "medium" + } + ] + }, + { + "type": "fact", + "insight": "System hardware and environment: The system uses a Raspberry Pi 3 controller with a Pi Camera module, running Raspbian as the base OS to enable operation and integration.", + "content": "3.3.1 Design of a multi-function activity monitoring and reporting system To design a multi-function activity monitoring and reporting system, Raspberry Pi 3 controller and Pi Camera module was adapted. Raspbian Operating system, being the most populous operating system for the raspberry pi controller was installed as the base operating system to enable swift operation and integration of the controller and the Pi camera module.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.1" + }, + { + "attribute": "source", + "value": "Page 28" + } + ] + }, + { + "type": "fact", + "insight": "Video processing pipeline begins with image acquisition from the Raspberry Pi Camera and converts the image to grayscale for processing.", + "content": "This entails using image processing techniques to process a video feed. This takes a succession of photographs in a frame, starting with image acquisition, which entails collecting the images from a source, in our case the raspberry pi camera module v2. The next step is to convert the image to grayscale so that it can be processed.", + "attributes": [ + { + "attribute": "section", + "value": "3.3" + } + ] + }, + { + "type": "fact", + "insight": "Motion detection is performed using a loaded Single-Shot Multibox Detection (SSD) MobileNet network accessed via Keras/OpenCV.", + "content": "Load Model Single-Shot multibox Detection (SSD) mobilenet network was utilized for this project. This was used to perform motion detection. The process involves loading the model using Keras/OpenCV to make it accessible for use.", + "attributes": [ + { + "attribute": "section", + "value": "3.3" + }, + { + "attribute": "method", + "value": "SSD MobileNet" + } + ] + }, + { + "type": "fact", + "insight": "Object detection and recognition are implemented with multibox detection to identify objects, draw bounding boxes, and perform localization and classification.", + "content": "iii. Object Detection and Recognition To detect multiple activities, object detection and recognition techniques is adapted. This uses systems like multibox detection to identify objects in the frame which draws a rectangle around its extent. This combines two tasks which are object localization and image classification. The initial is the process of identifying the object in the frame and drawing a frame around it, whereas the latter is to locate the presence of the located object and its types or classes.", + "attributes": [ + { + "attribute": "section", + "value": "3.3" + } + ] + }, + { + "type": "fact", + "insight": "Pose estimation uses a custom pre-trained MobileNet model to predict and track a person’s location by analyzing pose and orientation to recognize postural activities such as sitting and standing.", + "content": "iv. Pose Estimation A custom pre-trained mobilenet model was adapted for this section. This is a technique for predicting and tracking the location of a person by looking at the combination of the pose and the orientation of the person. This is used for recognizing postural activities like sitting and standing in this research.", + "attributes": [ + { + "attribute": "section", + "value": "3.3" + } + ] + }, + { + "type": "fact", + "insight": "Data and reports are stored in the cloud using Google Firebase services, enabling video, data storage, analysis, and notifications to the mobile interface and AI backend.", + "content": "v. Database and Storage In this section, data is collected, and reports are stored in the cloud using Google Firebase services. This keeps the videos, data and perform analysis on them. It is used to send notifications and request to the mobile interface and the A.I backend.", + "attributes": [ + { + "attribute": "section", + "value": "3.3" + } + ] + }, + { + "type": "fact", + "insight": "A mobile interface with rich UX supports monitoring and management of activities, allowing users to set tracked activities, timing, and report channels.", + "content": "vi. Mobile Interface A beautiful interface rich in User Experience (UX) was built for monitoring and managing activities. The application runs the multi-function activity monitoring and reporting system and give useful analytical information. It also allows the user to set activities to be tracked, the time and the report channel.", + "attributes": [ + { + "attribute": "section", + "value": "3.3" + } + ] + }, + { + "type": "comment", + "insight": "The page outlines a design-focused section that describes adapting hardware (Raspberry Pi 3 and camera) and software (Raspbian) to enable system operation, indicating a hardware-software integration focus.", + "content": "3.3.1 Design of a multi-function activity monitoring and reporting system To design a multi-function activity monitoring and reporting system, Raspberry Pi 3 controller and Pi Camera module was adapted. Raspbian Operating system, being the most populous operating system for the raspberry pi controller was installed as the base operating system to enable swift operation and integration of the controller and the Pi camera module. This section gives detailed information about data acquisition and preparation, video feed acquisition, activity 29", + "attributes": [ + { + "attribute": "section", + "value": "3.3.1" + } + ] + }, + { + "type": "fact", + "insight": "The section references detailed information about data acquisition, data preparation, and video feed acquisition, implying comprehensive coverage of data handling in this system.", + "content": "This section gives detailed information about data acquisition and preparation, video feed acquisition, activity", + "attributes": [ + { + "attribute": "section", + "value": "3.3" + } + ] + }, + { + "type": "fact", + "insight": "System uses Raspberry Pi 3 controller and Pi Camera module with Raspbian OS as the base to enable swift operation and integration.", + "content": "3.3.1 Design of a multi-function activity monitoring and reporting system To design a multi-function activity monitoring and reporting system, Raspberry Pi 3 controller and Pi Camera module was adapted. Raspbian Operating system, being the most populous operating system for the raspberry pi controller was installed as the base operating system to enable swift operation and integration of the controller and the Pi camera module.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.1 Design of a multi-function activity monitoring and reporting system" + }, + { + "attribute": "page", + "value": "29" + }, + { + "attribute": "device", + "value": "Raspberry Pi 3, Pi Camera" + }, + { + "attribute": "os", + "value": "Raspbian" + }, + { + "attribute": "source", + "value": "Page 29 text" + } + ] + }, + { + "type": "fact", + "insight": "Data acquisition and preparation involved collecting quality photos, removing noisy data, and augmenting data manually; more than a thousand samples were gathered.", + "content": "3.3.1 1 Data Acquisition and Preparation Data acquisition is the first step in the training process for every machine learning model. Even though the procedure may be automated in many ways, in this training scenario, quality photos were acquired, edited to remove noisy data, and augmented by manually collecting the data and preparing it. For training and testing, more than a thousand samples were gathered in this project.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.1.1 Data Acquisition and Preparation" + }, + { + "attribute": "page", + "value": "29" + }, + { + "attribute": "notes", + "value": ">1000 samples" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Video feed acquisition is from Pi Camera and can come from Homes or Hospitals depending on camera location; the video is processed frame by frame with computer vision techniques.", + "content": "3.3.1.2 Video Feed Acquisition This is the process of obtaining a video feed from the Pi Camera. This is the first step and one of the most crucial steps in any video utilized research. In this project work, video feeds are acquired from Homes or Hospitals depending on where the camera is mounted. The aim in this case is to get the video of the patient or an empty room. This video is processed frame by frame using computer vision techniques to enable analysis and processes to run. Below is a pseudo code of how video feed is acquired in this project. Step 1: START Step 2: Load an Image Step 3: Initialize a function 𝑉𝑖𝑑𝑒𝑜𝐶𝑎𝑝𝑡𝑢𝑟𝑒 ( ) 30", + "attributes": [ + { + "attribute": "section", + "value": "3.3.1.2 Video Feed Acquisition" + }, + { + "attribute": "page", + "value": "29" + }, + { + "attribute": "note", + "value": "Video feed from Pi Camera; Homes/Hospitals; pseudo code provided" + } + ] + }, + { + "type": "fact", + "insight": "Activity detection uses MobileNet/SSD with COCO dataset to detect >100 objects in real-time, but is limited to humans for this project; the custom MobileNet focuses on three classes: sitting, standing, and lying.", + "content": "3.3.1.3 Activity detection and recognition Mobilenets are one of the fastest machine learning models which are built, scaled and optimized mainly for low-level devices like the controller. Its design implements lightweight deep neural networks using proven depth-wise separable convolutions. The mobilenetSSD adapted in this project uses the coco dataset which has classes of over 100 objects. This can detect over 100 objects in realtime from our video feed but only limited to detecting and recognizing humans. The later, being the custom mobilenet uses a pre-trained model with only three classes: sitting, standing and laying as in fig.3.6.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.1.3 Activity detection and recognition" + }, + { + "attribute": "page", + "value": "29" + }, + { + "attribute": "note", + "value": "MobileNet/SSD; COCO; 100+ classes; three-class custom model" + } + ] + }, + { + "type": "fact", + "insight": "Figure 3.4 illustrates the circuit design for a multi-function activity monitoring and reporting system.", + "content": "Figure 3.4: Circuit design for multi-function activity monitoring and reporting system", + "attributes": [ + { + "attribute": "figure", + "value": "Figure 3.4" + }, + { + "attribute": "section", + "value": "3.3.1.2" + }, + { + "attribute": "source", + "value": "Page 30" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Data acquisition and preparation is the first step in training an ML model; the project collected quality photos, edited data to remove noise, and augmented data by manual collection; more than 1000 samples were gathered.", + "content": "3.3.1.1 Data Acquisition and Preparation Data acquisition is the first step in the training process for every machine learning model. Even though the procedure may be automated in many ways, in this training scenario, quality photos were acquired, edited to remove noisy data, and augmented by manually collecting the data and preparing it. For training and testing, more than a thousand samples were gathered in this project.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.1.1" + }, + { + "attribute": "data_volume", + "value": ">1000" + }, + { + "attribute": "data_cleaning", + "value": "edited to remove noisy data" + }, + { + "attribute": "augmentation", + "value": "manual collection" + }, + { + "attribute": "source", + "value": "Page 30" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Video feed acquisition is the process of obtaining a video feed from the Pi Camera and is a crucial step in video-based research.", + "content": "3.3.1.2 Video Feed Acquisition This is the process of obtaining a video feed from the Pi Camera. This is the first step and one of the crucial steps in any video utilized research. In this project work, video feeds are acquired from Homes or Hospitals depending on where the camera is mounted. The aim in this case is to get the video of the patient or an empty room. This video is processed frame by frame using computer vision techniques to enable analysis and processes to run. Below is a pseudo code of how video feed is acquired in this project. Step 1: START Step 2: Load an Image Step 3: Initialize a function 𝑉𝑖𝑑𝑒𝑜𝐶𝑎𝑝𝑡𝑢𝑟𝑒 ( )", + "attributes": [ + { + "attribute": "section", + "value": "3.3.1.2" + }, + { + "attribute": "hardware", + "value": "Pi Camera" + }, + { + "attribute": "video_goal", + "value": "video of patient or empty room" + }, + { + "attribute": "frame_processing", + "value": "frame-by-frame with computer vision techniques" + }, + { + "attribute": "source", + "value": "Page 30" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "opinion", + "insight": "The text states that video feed acquisition is 'one of the crucial steps' in video research, reflecting the authors' view on its importance.", + "content": "This is the first step and one of the crucial steps in any video utilized research.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.1.2" + }, + { + "attribute": "sentiment", + "value": "positive importance" + } + ] + }, + { + "type": "comment", + "insight": "The page explicitly mentions acquiring a video from Homes or Hospitals depending on camera mounting location and processing it frame by frame.", + "content": "In this project work, video feeds are acquired from Homes or Hospitals depending on where the camera is mounted. The aim in this case is to get the video of the patient or an empty room. This video is processed frame by frame using computer vision techniques to enable analysis and processes to run.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.1.2" + } + ] + }, + { + "type": "comment", + "insight": "The text presents a short pseudo code snippet: Step 1: START, Step 2: Load an Image, Step 3: Initialize a function VideoCapture().", + "content": "Below is a pseudo code of how video feed is acquired in this project. Step 1: START Step 2: Load an Image Step 3: Initialize a function 𝑉𝑖𝑑𝑒𝑜𝐶𝑎𝑝𝑡𝑢𝑟𝑒 ( )", + "attributes": [ + { + "attribute": "section", + "value": "3.3.1.2" + }, + { + "attribute": "pseudo_code", + "value": "Step 1 START; Step 2 Load an Image; Step 3 Initialize VideoCapture()" + } + ] + }, + { + "type": "comment", + "insight": "The page references Figure 3.4 and describes a pseudo code, indicating supplementary materials accompany the text.", + "content": "Figure 3.4: Circuit design for multi-function activity monitoring and reporting system", + "attributes": [ + { + "attribute": "section", + "value": "3.3.1" + } + ] + }, + { + "type": "fact", + "insight": "The page enumerates steps 4 through 9 of a video acquisition algorithm: initialize an infinite while loop; read each frame from the loop; perform operations and processes; display the video feed; press 'q' to exit; and STOP.", + "content": "Step 4: Initialize an infinite while loop Step 5: Read each frame from while loop Step 6: Perform operations and processes Step 7: Display video feed Step 8: Press “q” to exit Step 9: STOP", + "attributes": [ + { + "attribute": "section", + "value": "Video acquisition algorithm (Steps 4-9)" + }, + { + "attribute": "step_range", + "value": "4-9" + }, + { + "attribute": "page", + "value": "31" + }, + { + "attribute": "source", + "value": "Figure 3.5 (flowchart)" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "comment", + "insight": "The text notes that the flowchart shown below in figure 3.5 describes the algorithm for video acquisition.", + "content": "The flowchart shown below in figure 3.5 describes the algorithm for video acquisition", + "attributes": [ + { + "attribute": "section", + "value": "Video acquisition algorithm" + }, + { + "attribute": "figure", + "value": "Figure 3.5" + }, + { + "attribute": "page", + "value": "31" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "comment", + "insight": "Figure 3.5: Flowchart for video acquisition", + "content": "Figure 3.5: Flowchart for video acquisition", + "attributes": [ + { + "attribute": "section", + "value": "Figure caption" + }, + { + "attribute": "figure", + "value": "Figure 3.5" + }, + { + "attribute": "page", + "value": "31" + } + ] + }, + { + "type": "fact", + "insight": "MobileNetSSD outputs a scores tensor with shape (3, 300, 300) when a picture is input.", + "content": "A picture is entered into MobilenetSSD, which produces ( 3 , 300 , 300 ) scores.", + "attributes": [ + { + "attribute": "model", + "value": "MobileNet-SSD" + }, + { + "attribute": "output_shape", + "value": "(3,300,300)" + }, + { + "attribute": "section", + "value": "Architecture" + } + ] + }, + { + "type": "fact", + "insight": "There are 20 object categories, with 0 representing the backdrop in the scores.", + "content": "The 20 different item categories are represented by confidence levels in Scores, with 0 representing the backdrop.", + "attributes": [ + { + "attribute": "classes", + "value": "20" + }, + { + "attribute": "background_label", + "value": "0" + }, + { + "attribute": "section", + "value": "Architecture" + } + ] + }, + { + "type": "fact", + "insight": "SSD training requires only a ground truth box and an input picture.", + "content": "For each item during training, SSD just requires a ground truth box and an input picture.", + "attributes": [ + { + "attribute": "training_requirements", + "value": "ground_truth_box, input_picture" + }, + { + "attribute": "section", + "value": "Training" + } + ] + }, + { + "type": "fact", + "insight": "Default boxes are examined across various aspect ratios and feature maps.", + "content": "The small set of default boxes were examined (for example, 4) with various aspect ratios at each position in several feature maps with various sizes (for example, and in (b) and (c)).", + "attributes": [ + { + "attribute": "default_boxes_config", + "value": "aspect_ratios across multiple feature maps and sizes" + }, + { + "attribute": "section", + "value": "Architecture" + } + ] + }, + { + "type": "fact", + "insight": "Offsets and confidences are forecast for each default box for all object categories.", + "content": "The shape offsets and confidences for each 8 × 8 4 × 4default box for all object categories was forecast.", + "attributes": [ + { + "attribute": "prediction_scope", + "value": "offsets and confidences for each default box across all categories" + }, + { + "attribute": "section", + "value": "Architecture" + } + ] + }, + { + "type": "fact", + "insight": "During training, the first match of default boxes to ground truth defines positives; others are negatives.", + "content": "The first match of these ( ( 𝐶 1 , 𝐶 2 , \" . . , 𝐶 ) )default boxes to the ground truth boxes at training time.", + "attributes": [ + { + "attribute": "matching_criterion", + "value": "first_match_to_ground_truth_at_training_time" + }, + { + "attribute": "section", + "value": "Training" + } + ] + }, + { + "type": "fact", + "insight": "Two default boxes matched to two persons in a frame are considered positives; others are negatives.", + "content": "As an illustration in Figure 3.7, two default boxes with two persons in the frame is matched; these are considered positives, and the remaining default boxes are considered negatives.", + "attributes": [ + { + "attribute": "example_description", + "value": "two default boxes matched to two persons; positives; others negatives" + }, + { + "attribute": "section", + "value": "Training" + } + ] + }, + { + "type": "fact", + "insight": "The model loss is a weighted average of the localization loss and the confidence loss (Softmax).", + "content": "The model loss is a weighted average of the localization loss and the confidence loss (e.g. Softmax).", + "attributes": [ + { + "attribute": "loss_function", + "value": "weighted average of localization and confidence (Softmax)" + }, + { + "attribute": "section", + "value": "Training" + } + ] + }, + { + "type": "fact", + "insight": "Reporting channels for detected activities are SMS, Email, and In-App notifications.", + "content": "3.3.1.4 Reporting Channels To receive updates on activities detected and recognized, Short Message Service (SMS), Email and In-App notifications are used as channels for receiving information.", + "attributes": [ + { + "attribute": "reporting_channels", + "value": "SMS, Email, In-App" + }, + { + "attribute": "section", + "value": "3.3.1.4 Reporting Channels" + } + ] + }, + { + "type": "fact", + "insight": "SMS is sent via the Twilio API, and email/in-app via Nodemailer and push notifications.", + "content": "This uses Twilio Application Program Interface (API) to send SMS from the script that’s being executed on the controller. Nodemailer and Push notifications were adapted for email and in-app notification.", + "attributes": [ + { + "attribute": "sms_via", + "value": "Twilio API" + }, + { + "attribute": "email_via", + "value": "Nodemailer" + }, + { + "attribute": "in_app_via", + "value": "Push notifications" + }, + { + "attribute": "section", + "value": "3.3.1.4 Reporting Channels" + } + ] + }, + { + "type": "fact", + "insight": "The page mentions a flowchart (fig.3.8) illustrating a multi-function activity monitoring and reporting system.", + "content": "The flowchart shown below in fig.3.8 shows the multi-function activity monitoring and reporting system.", + "attributes": [ + { + "attribute": "figures", + "value": "Figure 3.7 and Figure 3.8 referenced; flowchart shown" + }, + { + "attribute": "section", + "value": "Figures/Illustrations" + } + ] + }, + { + "type": "comment", + "insight": "The page contains figure references (Figure 3.7, Figure 3.8) and describes an SSD framework and monitoring system.", + "content": "The flowchart shown below in fig.3.8 shows the multi-function activity monitoring and reporting system.", + "attributes": [ + { + "attribute": "section", + "value": "Figures/Illustrations" + } + ] + }, + { + "type": "fact", + "insight": "MobilenetSSD processes an input image to produce (3, 300, 300) scores.", + "content": "A picture is entered into MobilenetSSD, which produces ( 3 , 300 , 300 ) scores.", + "attributes": [ + { + "attribute": "model", + "value": "MobilenetSSD" + }, + { + "attribute": "context", + "value": "Page 33, Architecture for activity detection and recognition" + } + ] + }, + { + "type": "fact", + "insight": "Offsets from the default boxes are included in the predictions.", + "content": "Offset values from the default box ( 1 , 3000 , 4 𝑏𝑜𝑥𝑒𝑠 𝑎𝑛𝑑 ( 1 , 3000 , 21 ) ( 𝑐𝑥 , 𝑐𝑦 , 𝑤 , ℎ )are included in boxes.", + "attributes": [ + { + "attribute": "model", + "value": "default box offsets" + }, + { + "attribute": "context", + "value": "Page 33" + } + ] + }, + { + "type": "fact", + "insight": "There are 20 item categories represented by confidence levels, with 0 representing the backdrop.", + "content": "The 20 different item categories are represented by confidence levels in Scores, with 0 representing the backdrop.", + "attributes": [ + { + "attribute": "categories", + "value": "20" + }, + { + "attribute": "backdrop", + "value": "0" + } + ] + }, + { + "type": "fact", + "insight": "During training SSD requires a ground truth box and an input picture.", + "content": "For each item during training, SSD just requires a ground truth box and an input picture.", + "attributes": [ + { + "attribute": "process", + "value": "training" + } + ] + }, + { + "type": "fact", + "insight": "A small set of default boxes with various aspect ratios across multiple feature maps was examined.", + "content": "The small set of default boxes were examined (for example, 4) with various aspect ratios at each position in several feature maps with various sizes (for example, and in (b) and (c)).", + "attributes": [ + { + "attribute": "default_boxes", + "value": "various aspect ratios across multiple feature maps" + } + ] + }, + { + "type": "fact", + "insight": "The model forecasts shape offsets and confidences for each default box across all object categories.", + "content": "The shape offsets and confidences for each 8 × 8 4 × 4default box for all object categories was forecast.", + "attributes": [ + { + "attribute": "forecast", + "value": "shape offsets and confidences for all object categories" + } + ] + }, + { + "type": "fact", + "insight": "The first match between default boxes and ground truth boxes is considered positives; others are negatives.", + "content": "The first match of these ( ( 𝐶 1 , 𝐶 2 , \" . . , 𝐶 ) )default boxes to the ground truth boxes at training time. As an illustration in Figure 3.7, two default boxes with two persons in the frame is matched; these are considered positives, and the remaining default boxes are considered negatives.", + "attributes": [ + { + "attribute": "matching", + "value": "positive/negative assignment during training" + } + ] + }, + { + "type": "fact", + "insight": "The model loss is a weighted average of localization loss and confidence loss (e.g., Softmax).", + "content": "The model loss is a weighted average of the localization loss and the confidence loss (e.g. Softmax).", + "attributes": [ + { + "attribute": "loss_components", + "value": "localization and confidence (Softmax)" + } + ] + }, + { + "type": "fact", + "insight": "Figure 3.7 depicts the SSD Framework.", + "content": "Figure 3.7: Single Shot Detector (SSD) Framework", + "attributes": [ + { + "attribute": "figure", + "value": "3.7" + } + ] + }, + { + "type": "fact", + "insight": "Reporting channels for detected activities include SMS, Email, and In-App notifications.", + "content": "3.3.1.4 Reporting Channels To receive updates on activities detected and recognized, Short Message Service (SMS), Email and In-App notifications are used as channels for receiving information.", + "attributes": [ + { + "attribute": "channels", + "value": "SMS, Email, In-App" + } + ] + }, + { + "type": "fact", + "insight": "SMS is sent using the Twilio API.", + "content": "This uses Twilio Application Program Interface (API) to send SMS from the script that’s being executed on the controller.", + "attributes": [ + { + "attribute": "integration", + "value": "Twilio API" + } + ] + }, + { + "type": "fact", + "insight": "Nodemailer and Push notifications are used for email and in-app notifications.", + "content": "Nodemailer and Push notifications were adapted for email and in-app notification.", + "attributes": [ + { + "attribute": "integration", + "value": "Nodemailer; Push" + } + ] + }, + { + "type": "fact", + "insight": "A flowchart (Figure 3.8) illustrates the multi-function activity monitoring and reporting system.", + "content": "The flowchart shown below in fig.3.8 shows the multi-function activity monitoring and reporting system.", + "attributes": [ + { + "attribute": "figure", + "value": "3.8" + } + ] + }, + { + "type": "comment", + "insight": "The page references multiple figures (3.7 and 3.8) to illustrate the SSD framework and the reporting system.", + "content": "Figure 3.7: Single Shot Detector (SSD) Framework; The flowchart shown below in fig.3.8 shows the multi-function activity monitoring and reporting system.", + "attributes": [ + { + "attribute": "notes", + "value": "Figure references on page" + } + ] + }, + { + "type": "fact", + "insight": "Firebase Realtime Database is used to enable real-time data updates between the mobile interface and the controller, with data synced in real time and stored as JSON.", + "content": "To be able to update the data in realtime from the mobile interface to the controller and vice-versa, the Firebase Realtime Database was utilized, this is a database that is hosted in the cloud. Data is synced in real-time to the controller and saved as JSON. The Realtime Database instance is shared by the controller and the mobile interface which ensures that the device always have the most recent data available. Activities, camera information and tracking data are stored in the realtime database.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.2.1 Firebase Realtime Database" + }, + { + "attribute": "source", + "value": "Page 34-35" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Firebase Cloud Storage is used to store all videos recorded in the project and provides publicly accessible links for global access, with security for uploads/downloads and scalable storage.", + "content": "This project implements Cloud Storage for Firebase, a robust, user-friendly, and cost-efficient object storage solution created for Google scale. Regardless of network condition, the Firebase SDKs for Cloud Storage give Google security to file uploads and downloads for your Firebase apps. This stores all the videos recorded in the project and provide a publicly accessible link which can be accessed globally from anywhere in the world as in fig.3.9.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.2.2 Firebase Cloud Storage" + }, + { + "attribute": "source", + "value": "Page 34-35" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "comment", + "insight": "The document states that the mobile interface is fully developed for both iOS and Android.", + "content": "In this design, a fully fletched mobile interface for iOS and Android is built to enable managing of monitored and reported activities.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.3 Design of a mobile application for managing monitored and reported activities" + }, + { + "attribute": "source", + "value": "Page 34-35" + }, + { + "attribute": "confidence", + "value": "medium" + } + ] + }, + { + "type": "opinion", + "insight": "The text characterizes Expo as making mobile development setup easy and fast.", + "content": "Setting up for mobile development has been made easy and fast with expo. Expo is a React Native-based toolchain that assists in swiftly setting up applications.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.3.1 Expo" + }, + { + "attribute": "source", + "value": "Page 35" + }, + { + "attribute": "confidence", + "value": "medium" + } + ] + }, + { + "type": "fact", + "insight": "Expo provides libraries and features such as Barcode Scanner, MapView, and ImagePicker via the Expo SDK for local mobile functions.", + "content": "It also offers the Expo SDK, which can be used for some local mobile functions, like Barcode Scanner, MapView, ImagePicker, etc.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.3.1 Expo" + }, + { + "attribute": "source", + "value": "Page 35" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "React Native Chart Kit is used to visualize data with a stacked bar chart.", + "content": "Using a bar chart to visualize data in this project, a chart is a graphical representation of information. React Native Chart Kit is a popular chart library that aids in presenting data in an engaging way. It has multiple chart symbols like bar charts, pie charts, line charts, etc. In this project, a stacked bar chart was used as a way of displaying the visualized data.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.3.2 React Native Chart Kit" + }, + { + "attribute": "source", + "value": "Page 35" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "React Navigation is used to manage navigation between screens, with flow diagrams illustrating stack-like and tab-based navigation across Android and iOS.", + "content": "Rarely do mobile apps consist of only one screen. A navigator is often responsible for controlling how several displays are shown and switched between. A simple navigation solution is offered by React Navigation, which can display popular stack navigation and tabbed navigation patterns on both Android and iOS. fig.3.9 shows a flow diagram of React Navigation.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.3.3 React Navigation" + }, + { + "attribute": "source", + "value": "Page 35" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "comment", + "insight": "Figure 3.8 is referenced as the flowchart for the multi-function activity monitoring and reporting system on page 34.", + "content": "Figure 3.8: Flowchart for multi-function activity monitoring and reporting system", + "attributes": [ + { + "attribute": "section", + "value": "Figure 3.8 reference" + }, + { + "attribute": "source", + "value": "Page 34" + }, + { + "attribute": "confidence", + "value": "low" + } + ] + }, + { + "type": "fact", + "insight": "Figure 3.9 is referenced in the text as showing publicly accessible video links and is tied to the storage description.", + "content": "This stores all the videos recorded in the project and provide a publicly accessible link which can be accessed globally from anywhere in the world as in fig.3.9.", + "attributes": [ + { + "attribute": "section", + "value": "Figure 3.9 reference" + }, + { + "attribute": "source", + "value": "Page 34-35" + }, + { + "attribute": "confidence", + "value": "low" + } + ] + }, + { + "type": "fact", + "insight": "Firebase is used to design a centralized database architecture with a storage facility for multiple activity tracking and reporting system.", + "content": "Firebase services being known for its ease of use and scalability was adapted for this project to design a centralized database with a storage facility.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.2" + }, + { + "attribute": "topic", + "value": "centralized database architecture with storage facility" + }, + { + "attribute": "source", + "value": "Page 35" + } + ] + }, + { + "type": "fact", + "insight": "Firebase Realtime Database provides real-time data updates between the mobile interface and the controller, with data stored as JSON.", + "content": "To be able to update the data in realtime from the mobile interface to the controller and vice-versa, the Firebase Realtime Database was utilized, this is a database that is hosted in the cloud. Data is synced in real-time to the controller and saved as JSON. The Realtime Database instance is shared by the controller and the mobile interface which ensures that the device always have the most recent data available. Activities, camera information and tracking data are stored in the realtime database.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.2.1" + }, + { + "attribute": "data_format", + "value": "JSON" + }, + { + "attribute": "data_type", + "value": "Realtime Database" + }, + { + "attribute": "source", + "value": "Page 35" + } + ] + }, + { + "type": "fact", + "insight": "Firebase Cloud Storage is used for object storage; it is robust, user-friendly, and cost-efficient, with SDKs providing security for uploads and downloads.", + "content": "This project implements Cloud Storage for Firebase, a robust, user-friendly, and cost-efficient object storage solution created for Google scale. Regardless of network condition, the Firebase SDKs for Cloud Storage give Google security to file uploads and downloads for your Firebase apps.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.2.2" + }, + { + "attribute": "storage_type", + "value": "Cloud Storage" + }, + { + "attribute": "security", + "value": "Google security for uploads/downloads" + }, + { + "attribute": "source", + "value": "Page 35" + } + ] + }, + { + "type": "fact", + "insight": "Videos recorded in the project are stored in Firebase Cloud Storage and linked via publicly accessible links.", + "content": "This stores all the videos recorded in the project and provide a publicly accessible link which can be accessed globally from anywhere in the world as in fig.3.9.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.2.2" + }, + { + "attribute": "topic", + "value": "video storage and public links" + }, + { + "attribute": "source", + "value": "Page 35" + } + ] + }, + { + "type": "fact", + "insight": "The mobile application is designed as a fully fledged interface for both iOS and Android to manage monitored and reported activities.", + "content": "In this design, a fully fletched mobile interface for iOS and Android is built to enable managing of monitored and reported activities.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.3" + }, + { + "attribute": "topic", + "value": "mobile application design" + }, + { + "attribute": "source", + "value": "Page 35" + } + ] + }, + { + "type": "comment", + "insight": "Figure 3.8 is captioned as a flowchart for a multi-function activity monitoring and reporting system.", + "content": "Figure 3.8: Flowchart for multi-function activity monitoring and reporting system", + "attributes": [ + { + "attribute": "figure", + "value": "3.8" + }, + { + "attribute": "type", + "value": "caption" + }, + { + "attribute": "source", + "value": "Page 35" + } + ] + }, + { + "type": "fact", + "insight": "Expo is a React Native-based toolchain that assists in swiftly setting up applications.", + "content": "Expo is a React Native-based toolchain that assists in swiftly setting up applications.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.3.1 Expo" + }, + { + "attribute": "page", + "value": "36" + }, + { + "attribute": "source", + "value": "Document" + }, + { + "attribute": "confidence", + "value": "0.95" + } + ] + }, + { + "type": "fact", + "insight": "Expo provides the user interface and service components and a variety of tools for creating and testing Native React apps.", + "content": "It provides the user interface and service components that are generally present in third-party Native React Native components as well as a variety of tools for the process of creating and testing Native React apps.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.3.1 Expo" + }, + { + "attribute": "page", + "value": "36" + }, + { + "attribute": "source", + "value": "Document" + }, + { + "attribute": "confidence", + "value": "0.9" + } + ] + }, + { + "type": "fact", + "insight": "Expo SDK includes functions for local mobile features such as Barcode Scanner, MapView, and ImagePicker.", + "content": "It also offers the Expo SDK, which can be used for some local mobile functions, like Barcode Scanner, MapView, ImagePicker, etc.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.3.1 Expo" + }, + { + "attribute": "page", + "value": "36" + }, + { + "attribute": "source", + "value": "Document" + }, + { + "attribute": "confidence", + "value": "0.9" + } + ] + }, + { + "type": "fact", + "insight": "In this project, Expo is adapted as the framework to kickstart the mobile interface.", + "content": "In this project, it’s adapted as the framework to kickstart the mobile interface.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.3.1 Expo" + }, + { + "attribute": "page", + "value": "36" + }, + { + "attribute": "source", + "value": "Document" + }, + { + "attribute": "confidence", + "value": "0.9" + } + ] + }, + { + "type": "fact", + "insight": "React Native Chart Kit is a chart library used to present data; it supports multiple chart types such as bar, pie, and line charts.", + "content": "React Native Chart Kit is a popular chart library that aids in presenting data in an engaging way. It has multiple chart symbols like bar charts, pie charts, line charts, e.t.c.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.3.2 React Native Chart Kit" + }, + { + "attribute": "page", + "value": "36" + }, + { + "attribute": "source", + "value": "Document" + }, + { + "attribute": "confidence", + "value": "0.9" + } + ] + }, + { + "type": "fact", + "insight": "A stacked bar chart was used in this project to display the visualized data.", + "content": "In this project, a stacked bar chart was used as a way of displaying the visualized data.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.3.2 React Native Chart Kit" + }, + { + "attribute": "page", + "value": "36" + }, + { + "attribute": "source", + "value": "Document" + }, + { + "attribute": "confidence", + "value": "0.9" + } + ] + }, + { + "type": "fact", + "insight": "React Navigation provides a simple navigation solution supporting stack and tabbed patterns on Android and iOS.", + "content": "A simple navigation solution is offered by React Navigation, which can display popular stack navigation and tabbed navigation patterns on both Android and iOS.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.3.3 React Navigation" + }, + { + "attribute": "page", + "value": "36" + }, + { + "attribute": "source", + "value": "Document" + }, + { + "attribute": "confidence", + "value": "0.9" + } + ] + }, + { + "type": "comment", + "insight": "fig.3.9 shows a flow diagram of React Navigation.", + "content": "fig.3.9 shows a flow diagram of React Navigation.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.3.3 React Navigation" + }, + { + "attribute": "page", + "value": "36" + }, + { + "attribute": "source", + "value": "Document" + }, + { + "attribute": "confidence", + "value": "0.9" + } + ] + }, + { + "type": "opinion", + "insight": "Rarely do mobile apps consist of only one screen.", + "content": "Rarely do mobile apps consist of only one screen.", + "attributes": [ + { + "attribute": "section", + "value": "3.3.3.3 React Navigation" + }, + { + "attribute": "page", + "value": "36" + }, + { + "attribute": "source", + "value": "Document" + }, + { + "attribute": "confidence", + "value": "0.8" + } + ] + }, + { + "type": "fact", + "insight": "BEME asserts the design was not cost-effective due to inflation and high material costs.", + "content": "3.4 Bill of Engineering Measurement and Evaluation (BEME) The design of multi-function activity reporting and monitoring system was not cost-effective due to the inflation and high cost of materials.", + "attributes": [ + { + "attribute": "section", + "value": "3.4 BEME" + } + ] + }, + { + "type": "fact", + "insight": "Table 3.1 lists the bill for the multi-function activity reporting and monitoring system including component costs and total.", + "content": "Table 3.1: The bill for the multi-function activity reporting and monitoring system S/N Component Description Quantity Cost ( N ) 1 Raspberry Pi Model 3 1 100,000 2 Pi Camera 1080p HD Webcam 5MP OV5647 Sensor 1 10,000 Total 110,000", + "attributes": [ + { + "attribute": "section", + "value": "Table 3.1; 3.4" + } + ] + }, + { + "type": "fact", + "insight": "Major Research Methods describe activity detection and recognition using AI on digital cameras/video feed, with SSD for bounding boxes and Mobilenet for manual samples.", + "content": "3.5 Major Research Methods and its Findings In a view of considering activities detection and recognition, activities were obtained by the digital cameras and the video feed which was processed to get the activity information using Artificial Intelligence. Bounding boxes on this information were located using Single Shot Detection system (SSD). Manual samples from health and security were acquired using mobilenets.", + "attributes": [ + { + "attribute": "section", + "value": "3.5 Major Research Methods and its Findings" + } + ] + }, + { + "type": "comment", + "insight": "Figure 3.9 caption indicates an Illustration of React Navigation on this page.", + "content": "Figure 3.9: Illustration of React Navigation", + "attributes": [ + { + "attribute": "section", + "value": "Figure caption" + } + ] + }, + { + "type": "fact", + "insight": "Bounding boxes are located using the Single Shot Detection system (SSD).", + "content": "Bounding boxes on this information were located using Single Shot Detection system (SSD).", + "attributes": [ + { + "attribute": "method", + "value": "SSD" + } + ] + }, + { + "type": "fact", + "insight": "Manual samples from health and security were acquired using MobileNet (mobilenets).", + "content": "Manual samples from health and security were acquired using mobilenets.", + "attributes": [ + { + "attribute": "technology", + "value": "MobileNet" + } + ] + }, + { + "type": "fact", + "insight": "Chapter 4 presents results and discussion by evaluating the implemented methodology from Chapter 3 to assess system performance and quality of service.", + "content": "This chapter discusses and evaluates the results obtained from the implemented methodology of Chapter three. Given the initial stated parameters, these serves to evaluate the performance of the system and quality of service of the system.", + "attributes": [ + { + "attribute": "section", + "value": "4.0 Overview" + }, + { + "attribute": "chapter", + "value": "Chapter 4" + } + ] + }, + { + "type": "fact", + "insight": "Diverse testing locations were used in Keffi, Abuja, and Minna.", + "content": "Diverse locations were used within Keffi, Abuja and Minna for the model testing.", + "attributes": [ + { + "attribute": "section", + "value": "4.0 Overview" + }, + { + "attribute": "location_set", + "value": "Keffi, Abuja, Minna" + } + ] + }, + { + "type": "fact", + "insight": "Images and real-time video feeds were used to test the activity-detection algorithm.", + "content": "Images and real-time video feeds were used to test the algorithm to detect activities.", + "attributes": [ + { + "attribute": "section", + "value": "4.2" + }, + { + "attribute": "test_mode", + "value": "real-time video feeds" + } + ] + }, + { + "type": "fact", + "insight": "The test videos included a hospital room scene and a home room to assess model performance from different angles.", + "content": "The video feeds being a hospital room scene and a home room, it was prepared to test how intelligent the model was from every angle.", + "attributes": [ + { + "attribute": "section", + "value": "4.2" + }, + { + "attribute": "test_scenes", + "value": "hospital room, home room" + } + ] + }, + { + "type": "fact", + "insight": "Camera placement (top and center) affected accuracy levels across results.", + "content": "Camera was placed at the top and center of the room and each result gave different accuracy level.", + "attributes": [ + { + "attribute": "section", + "value": "4.2" + }, + { + "attribute": "camera_position", + "value": "top and center" + } + ] + }, + { + "type": "fact", + "insight": "For motion detection, accuracy was roughly similar across different camera locations, while center placement yielded better results for other activities.", + "content": "As for the motion detection, accuracy level was almost the same at multiple camera location but for the other activities, positioning the camera at the center produced a better and more accurate prediction.", + "attributes": [ + { + "attribute": "section", + "value": "4.2" + }, + { + "attribute": "analysis", + "value": "motion_detection_vs_other_activities" + } + ] + }, + { + "type": "fact", + "insight": "The implementation used a Raspberry Pi controller and a 1080P Pi Camera, with results documented in Table 4.1.", + "content": "This work was implemented using a Raspberry Pi controller and a 1080P Pi Camera and the result was documented as in Table 4.1.", + "attributes": [ + { + "attribute": "section", + "value": "4.2" + }, + { + "attribute": "hardware", + "value": "Raspberry Pi + 1080P Pi Camera" + } + ] + }, + { + "type": "fact", + "insight": "Table 4.1 reports specific accuracy figures for different locations and activities.", + "content": "Table 4.1: Performance of the A.I at different positions in a room Location Position Activities Success Rate Hospital Top Motion 90% Home Center Motion 95% Hospital Top Posture 45% Home Center Posture 70%", + "attributes": [ + { + "attribute": "section", + "value": "4.2" + }, + { + "attribute": "table", + "value": "Table 4.1" + } + ] + }, + { + "type": "fact", + "insight": "The Raspberry Pi is designed to operate at 5 volts with a tolerance of 5% (4.75–5.25 V), and will not turn on with voltage below that.", + "content": "The Raspberry Pi is designed to operate at 5 volts with a tolerance of 5%. (4.75 - 5.25 volts). The Pi won't turn on if you give less voltage than is necessary.", + "attributes": [ + { + "attribute": "section", + "value": "4.2" + }, + { + "attribute": "power_spec", + "value": "5V ±5%" + } + ] + }, + { + "type": "fact", + "insight": "The Pi can receive up to 30 frames per second from the Pi Camera, which is processed by the software.", + "content": "It receives up to 30 fps feed from the Pi Camera which is processed by the software part of the project.", + "attributes": [ + { + "attribute": "section", + "value": "4.2" + }, + { + "attribute": "fps", + "value": "up to 30 fps" + } + ] + }, + { + "type": "comment", + "insight": "The text frames the evaluation as testing the model’s intelligence from every angle.", + "content": "to test how intelligent the model was from every angle.", + "attributes": [ + { + "attribute": "section", + "value": "4.2" + }, + { + "attribute": "note", + "value": "model_intelligence_evaluation" + } + ] + }, + { + "type": "fact", + "insight": "The testing shows that motion detection performance can be high (up to 95%) depending on camera position.", + "content": "Home Center Motion 95%", + "attributes": [ + { + "attribute": "section", + "value": "4.2" + }, + { + "attribute": "table_entry", + "value": "Home Center Motion 95%" + } + ] + }, + { + "type": "fact", + "insight": "Motion detection draws a rectangle around the detected object (a human) using the Raspberry Pi Camera, with detection controlled by a time interval set by the user via the mobile interface and only when the interval is within the current time.", + "content": "The motion detection system as shown in fig.4.1 and fig.4.2. This is what the Raspberry Pi Camera sees. It detects and draws a rectangle around the detected object, in this case, a human. The detection works with time set by the user of the mobile interface. Motion is being detected only when the time interval set is within the current time.", + "attributes": [ + { + "attribute": "section", + "value": "4.2.1 Motion Detection" + }, + { + "attribute": "camera", + "value": "Raspberry Pi Camera" + }, + { + "attribute": "figures", + "value": "fig.4.1 and fig.4.2" + }, + { + "attribute": "time_control", + "value": "time interval set by user via mobile interface" + }, + { + "attribute": "restriction", + "value": "detection only when interval is within current time" + } + ] + }, + { + "type": "fact", + "insight": "Posture recognition detects activities such as lying down, sitting and standing; it performs best when the whole body is captured within the camera, and accuracy improves notably when the face is visible to the camera.", + "content": "4.2.2 Posture Recognition To detect activities such as lying down, sitting and standing, posture recognition was used as described in the use-case and flow chart diagrams in chapter 3. This performs well if the whole body of the person is captured within the camera. It’s also noticed that accuracy skyrocketed if the face of the person is visible to the camera which serves as a good guide to other key points. This is shown in fig.4.3.", + "attributes": [ + { + "attribute": "section", + "value": "4.2.2 Posture Recognition" + }, + { + "attribute": "condition_body", + "value": "whole body captured within the camera" + }, + { + "attribute": "face_visibility_effect", + "value": "accuracy skyrocketed if the face is visible" + }, + { + "attribute": "reference_figure", + "value": "fig.4.3" + } + ] + }, + { + "type": "comment", + "insight": "Figure 4.3 caption notes Body Posture Detection of a person sitting down.", + "content": "Figure 4.3: Body Posture Detection of a person sitting down", + "attributes": [ + { + "attribute": "section", + "value": "4.2.2 Posture Recognition" + }, + { + "attribute": "figure", + "value": "fig.4.3" + } + ] + }, + { + "type": "fact", + "insight": "Realtime communication was needed for a swift synchronization between the hardware prototype and the mobile interface.", + "content": "Realtime communication was needed for a swift synchronization between the hardware prototype and the mobile interface.", + "attributes": [ + { + "attribute": "section", + "value": "4.3 Results and Discussion" + }, + { + "attribute": "topic", + "value": "real-time sync between hardware and mobile interface" + } + ] + }, + { + "type": "fact", + "insight": "Firebase Realtime Database was tested to achieve this with a database schema that tracks activities created and monitored.", + "content": "Firebase realtime database was tested to achieve this with a database schema that tracks activities created and monitored.", + "attributes": [ + { + "attribute": "section", + "value": "4.3 Results and Discussion" + }, + { + "attribute": "database", + "value": "Firebase Realtime Database" + }, + { + "attribute": "schema", + "value": "tracks activities created and monitored" + } + ] + }, + { + "type": "fact", + "insight": "With a 1 GB storage limit in Cloud Firebase Storage, videos of recorded tracked activities were securely stored on the cloud.", + "content": "With 1GB storage limit capacity in the Cloud Firebase Storage, videos of recorded tracked activities were securely stored on the cloud.", + "attributes": [ + { + "attribute": "section", + "value": "4.3 Results and Discussion" + }, + { + "attribute": "storage", + "value": "1GB limit in Cloud Firebase Storage" + }, + { + "attribute": "security", + "value": "videos securely stored on the cloud" + } + ] + }, + { + "type": "fact", + "insight": "Posture recognition was used to detect activities such as lying down, sitting, and standing, as described in the use-case and flow chart diagrams in chapter 3.", + "content": "To detect activities such as lying down, sitting and standing, posture recognition was used as described in the use-case and flow chart diagrams in chapter 3.", + "attributes": [ + { + "attribute": "section", + "value": "4.2.2 Posture Recognition" + }, + { + "attribute": "source", + "value": "Page 40" + }, + { + "attribute": "related_section", + "value": "use-case and flow chart diagrams in chapter 3" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Posture recognition performance is better when the whole body is captured by the camera, and accuracy increases when the face is visible, guiding other key points (as shown in Fig. 4.3).", + "content": "This performs well if the whole body of the person is captured within the camera. It’s also noticed that accuracy skyrocketed if the face of the person is visible to the camera which serves as a good guide to other key points. This is shown in fig.4.3.", + "attributes": [ + { + "attribute": "section", + "value": "4.2.2 Posture Recognition" + }, + { + "attribute": "figure", + "value": "Fig. 4.3" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "comment", + "insight": "Figure 4.3 caption is provided to illustrate body posture detection of a person sitting down.", + "content": "Figure 4.3: Body Posture Detection of a person sitting down", + "attributes": [ + { + "attribute": "section", + "value": "4.2.2 Posture Recognition" + }, + { + "attribute": "figure", + "value": "4.3" + }, + { + "attribute": "note", + "value": "Caption/visual reference" + } + ] + }, + { + "type": "comment", + "insight": "The section header 4.3 introduces results and discussion for a centralized database architecture with a storage scheme for multiple activity tracking and reporting system.", + "content": "4.3 Results and Discussion for centralized database architecture with storage scheme", + "attributes": [ + { + "attribute": "section", + "value": "4.3" + }, + { + "attribute": "confidence", + "value": "medium" + } + ] + }, + { + "type": "fact", + "insight": "Real-time communication was needed for swift synchronization between the hardware prototype and the mobile interface.", + "content": "Realtime communication was needed for a swift synchronization between the hardware prototype and the mobile interface.", + "attributes": [ + { + "attribute": "section", + "value": "4.3" + }, + { + "attribute": "topic", + "value": "Real-time synchronization" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Firebase realtime database was tested to achieve real-time synchronization with a database schema that tracks activities created and monitored.", + "content": "Firebase realtime database was tested to achieve this with a database schema that tracks activities created and monitored.", + "attributes": [ + { + "attribute": "section", + "value": "4.3" + }, + { + "attribute": "technology", + "value": "Firebase Realtime Database" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "With a 1 GB storage limit in Cloud Firebase Storage, videos of recorded tracked activities were securely stored on the cloud.", + "content": "With 1GB storage limit capacity in the Cloud Firebase Storage, videos of recorded tracked activities were securely stored on the cloud.", + "attributes": [ + { + "attribute": "section", + "value": "4.3" + }, + { + "attribute": "storage", + "value": "Cloud Firebase Storage 1 GB limit" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "The project uses a simple structure schema for the Firebase Realtime Database.", + "content": "This project utilized a simple structure schema for the realtime database.", + "attributes": [ + { + "attribute": "section", + "value": "Firebase Realtime Database (4.3.1)" + }, + { + "attribute": "page", + "value": "41" + }, + { + "attribute": "source", + "value": "Page 41" + } + ] + }, + { + "type": "fact", + "insight": "An activities collection stores activities created by a particular user.", + "content": "An activities collection which has the activities that are created by a certain user as shown in fig.4.5.", + "attributes": [ + { + "attribute": "collection", + "value": "activities" + }, + { + "attribute": "purpose", + "value": "store activities created by a user" + }, + { + "attribute": "figure", + "value": "Fig.4.5" + }, + { + "attribute": "page", + "value": "41" + } + ] + }, + { + "type": "fact", + "insight": "A camera collection stores all active camera data and the activities the camera is supposed to track.", + "content": "The camera collection stores all active camera data with the activities the camera is supposed to track as shown in fig.4.6.", + "attributes": [ + { + "attribute": "collection", + "value": "camera" + }, + { + "attribute": "purpose", + "value": "store active camera data and associated tracked activities" + }, + { + "attribute": "figure", + "value": "Fig.4.6" + }, + { + "attribute": "page", + "value": "41" + } + ] + }, + { + "type": "fact", + "insight": "A tracked collection serves as a notification collection, storing detected activity information.", + "content": "The tracked collection serves as a notification collection, this stores the activity information that is detected as shown in fig.4.7.", + "attributes": [ + { + "attribute": "collection", + "value": "tracked" + }, + { + "attribute": "purpose", + "value": "store detected activity information as notifications" + }, + { + "attribute": "figure", + "value": "Fig.4.7" + }, + { + "attribute": "page", + "value": "41" + } + ] + }, + { + "type": "comment", + "insight": "Figure references (Fig.4.5, Fig.4.6, Fig.4.7) illustrate the respective collection tables in Firebase.", + "content": "Figure 4.5: Activities collection table on Firebase Figure 4.6: Camera collection table on Firebase Figure 4.7: Tracked collection table on Firebase", + "attributes": [ + { + "attribute": "figures", + "value": "Fig.4.5, Fig.4.6, Fig.4.7" + }, + { + "attribute": "page", + "value": "41" + }, + { + "attribute": "section", + "value": "Firebase Realtime Database (4.3.1)" + } + ] + }, + { + "type": "fact", + "insight": "The project utilizes a Firebase Realtime Database with a simple structure schema.", + "content": "This project utilized a simple structure schema for the realtime database.", + "attributes": [ + { + "attribute": "section", + "value": "4.3.1 Firebase Realtime database" + }, + { + "attribute": "page", + "value": "42" + }, + { + "attribute": "source", + "value": "Document" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "There is an activities collection that stores activities created by a user.", + "content": "An activities collection which has the activities that are created by a certain user as shown in fig.4.5.", + "attributes": [ + { + "attribute": "section", + "value": "4.3.1 Firebase Realtime database" + }, + { + "attribute": "page", + "value": "42" + }, + { + "attribute": "source", + "value": "Document" + }, + { + "attribute": "confidence", + "value": "high" + }, + { + "attribute": "figures", + "value": "fig.4.5" + } + ] + }, + { + "type": "fact", + "insight": "There is a camera collection that stores all active camera data and the activities the camera is supposed to track.", + "content": "The camera collection stores all active camera data with the activities the camera is supposed to track as shown in fig.4.6.", + "attributes": [ + { + "attribute": "section", + "value": "4.3.1 Firebase Realtime database" + }, + { + "attribute": "page", + "value": "42" + }, + { + "attribute": "source", + "value": "Document" + }, + { + "attribute": "confidence", + "value": "high" + }, + { + "attribute": "figures", + "value": "fig.4.6" + } + ] + }, + { + "type": "fact", + "insight": "There is a tracked collection that serves as a notification collection, storing the activity information that is detected.", + "content": "The tracked collection serves as a notification collection, this stores the activity information that is detected as shown in fig.4.7.", + "attributes": [ + { + "attribute": "section", + "value": "4.3.1 Firebase Realtime database" + }, + { + "attribute": "page", + "value": "42" + }, + { + "attribute": "source", + "value": "Document" + }, + { + "attribute": "confidence", + "value": "high" + }, + { + "attribute": "figures", + "value": "fig.4.7" + } + ] + }, + { + "type": "fact", + "insight": "Figure references indicate the corresponding collections (Activities, Camera, Tracked) on Firebase.", + "content": "Figure 4.5: Activities collection table on Firebase Figure 4.6: Camera collection table on Firebase 42", + "attributes": [ + { + "attribute": "section", + "value": "4.3.1 Firebase Realtime database" + }, + { + "attribute": "page", + "value": "42" + }, + { + "attribute": "source", + "value": "Document" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Introduction screen welcomes the user to the application and provides a CTA for login, registration, or password retrieval.", + "content": "The introduction screen welcomes user to the application as seen in fig.4.8, this has a good user experience (UX) as the user understands briefly what the application is built to do. It also provides call to action (CTA) that allows the user to login, register or retrieve their password.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.1 Introduction Screen" + }, + { + "attribute": "page", + "value": "43" + }, + { + "attribute": "figure", + "value": "Fig.4.8" + }, + { + "attribute": "topic", + "value": "Introduction Screen" + }, + { + "attribute": "source", + "value": "Mobile application 4.x (section 4.4)" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "opinion", + "insight": "The text asserts that the introduction screen provides good UX.", + "content": "this has a good user experience (UX) as the user understands briefly what the application is built to do.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.1 Introduction Screen" + }, + { + "attribute": "page", + "value": "43" + }, + { + "attribute": "sentiment", + "value": "positive" + }, + { + "attribute": "source", + "value": "Page 43, Introduction Screen" + }, + { + "attribute": "confidence", + "value": "medium" + } + ] + }, + { + "type": "fact", + "insight": "The introduction screen includes a call to action allowing login, registration, or password retrieval.", + "content": "It also provides call to action (CTA) that allows the user to login, register or retrieve their password.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.1 Introduction Screen" + }, + { + "attribute": "page", + "value": "43" + }, + { + "attribute": "figure", + "value": "Fig.4.8" + }, + { + "attribute": "source", + "value": "Introduction Screen text" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Login screen requires authentication to access monitoring and reporting aspects of the system; authentication is performed by the login screen.", + "content": "To be able to access the monitoring and reporting aspects of the system, an authentication is needed which is performed by the login screen. A user needs to validate he has access to this application before being granted access. Hence, the interface shown in fig.4.9 handles the authentication process.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.2 Login Screen" + }, + { + "attribute": "page", + "value": "43" + }, + { + "attribute": "figure", + "value": "Fig.4.9" + }, + { + "attribute": "topic", + "value": "Authentication" + }, + { + "attribute": "source", + "value": "Page 43, 4.4.2" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Registration screen allows new users to create an account to access the system; it asks a series of details for validation and verification; registration mechanism shown in fig.5.0.", + "content": "This allows new users create an account to be able to access the system. This asks series of details to get user information for validation and verification. The screen shown below in fig.5.0 handles the registration mechanism.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.3 Registration Screen" + }, + { + "attribute": "page", + "value": "44" + }, + { + "attribute": "figure", + "value": "Fig.5.0" + }, + { + "attribute": "source", + "value": "Registration Screen text" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "comment", + "insight": "Figure references show fig.4.8 for Introduction Screen and fig.4.9 for Login Screen; potential inconsistency in page numbering with fig.5.0 mention on the same page.", + "content": "Figure 4.8: Introduction Screen and Figure 4.9: Login Screen are shown; text for registration mentions fig.5.0, which may indicate a mismatch or typographical error.", + "attributes": [ + { + "attribute": "section", + "value": "4.4 (Screens)" + }, + { + "attribute": "page", + "value": "43-44" + }, + { + "attribute": "notes", + "value": "Possible figure-numbering inconsistency (fig.5.0 vs fig.4.x)" + } + ] + }, + { + "type": "fact", + "insight": "The Introduction Screen provides a call to action (CTA) that allows the user to login, register or retrieve their password.", + "content": "The introduction screen welcomes user to the application as seen in fig.4.8, this has a good user experience (UX) as the user understands briefly what the application is built to do. It also provides call to action (CTA) that allows the user to login, register or retrieve their password.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.1 Introduction Screen" + }, + { + "attribute": "page", + "value": "44" + }, + { + "attribute": "figure", + "value": "Fig.4.8" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "opinion", + "insight": "The Introduction Screen is described as having a good user experience (UX).", + "content": "The introduction screen welcomes user to the application as seen in fig.4.8, this has a good user experience (UX) as the user understands briefly what the application is built to do.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.1 Introduction Screen" + }, + { + "attribute": "page", + "value": "44" + }, + { + "attribute": "figure", + "value": "Fig.4.8" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Authentication is required to access the monitoring and reporting aspects of the system, and the login screen handles the authentication process.", + "content": "To be able to access the monitoring and reporting aspects of the system, an authentication is needed which is performed by the login screen.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.2 Login Screen" + }, + { + "attribute": "page", + "value": "44" + }, + { + "attribute": "figure", + "value": "Fig.4.9" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "A user must validate that they have access to this application before being granted access.", + "content": "A user needs to validate he has access to this application before being granted access.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.2 Login Screen" + }, + { + "attribute": "page", + "value": "44" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "comment", + "insight": "The page references Figures 4.8 and 4.9 to illustrate the Introduction and Login screens.", + "content": "Figure 4.8: Introduction Screen", + "attributes": [ + { + "attribute": "section", + "value": "4.4.1 Introduction Screen" + }, + { + "attribute": "page", + "value": "44" + }, + { + "attribute": "figure", + "value": "Fig.4.8" + } + ] + }, + { + "type": "fact", + "insight": "Registration Screen is presented (Figure 4.10).", + "content": "Figure 4.10: Registration Screen", + "attributes": [ + { + "attribute": "section", + "value": "4.4.x Screens (Registration)" + }, + { + "attribute": "page", + "value": "45" + }, + { + "attribute": "figure", + "value": "4.10" + } + ] + }, + { + "type": "fact", + "insight": "Forgot Password Screen describes a user interface to reset lost login credentials to regain access.", + "content": "4.4.4 Forget Password Screen This forget password screen is shown in fig.4.11 serves as an interface where users can reset their lost login credentials in order to regain access to the application.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.4 Forget Password Screen" + }, + { + "attribute": "page", + "value": "45" + }, + { + "attribute": "figure", + "value": "4.11" + } + ] + }, + { + "type": "fact", + "insight": "Forgot Password Screen is labeled as Figure 4.11 in the document.", + "content": "Figure 4.11: Forgot Password Screen", + "attributes": [ + { + "attribute": "section", + "value": "4.4.4 Forget Password Screen" + }, + { + "attribute": "page", + "value": "46" + }, + { + "attribute": "figure", + "value": "4.11" + } + ] + }, + { + "type": "fact", + "insight": "Home Screen is the main interface and acts as a connector to all other screens.", + "content": "4.4.5 Home Screen This serves as the main interface for this application. It serves as a connector to all the other screens.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.5 Home Screen" + }, + { + "attribute": "page", + "value": "46" + } + ] + }, + { + "type": "fact", + "insight": "Home Screen provides access to activities and reports with visual statistical data such as a bar chart (fig.4.12).", + "content": "Users can see activities and reports with visual statistical data such as the bar chart as seen in fig.4.12.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.5 Home Screen" + }, + { + "attribute": "page", + "value": "46" + }, + { + "attribute": "visualization", + "value": "bar chart" + } + ] + }, + { + "type": "fact", + "insight": "Home Screen displays recent activities with time and percentage.", + "content": "It also displays recent activities reported with the time and percentage.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.5 Home Screen" + }, + { + "attribute": "page", + "value": "46" + }, + { + "attribute": "aspect", + "value": "recent activities with time and percentage" + } + ] + }, + { + "type": "fact", + "insight": "Access to the Home Screen is restricted to authenticated users who passed login validation.", + "content": "This page is only shown to authenticated users, i.e., those who have passed the login validation.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.5 Home Screen" + }, + { + "attribute": "page", + "value": "46" + }, + { + "attribute": "security", + "value": "authenticated users only" + } + ] + }, + { + "type": "fact", + "insight": "Home Screen is the main interface and a connector to all other screens.", + "content": "4.4.5 Home Screen This serves as the main interface for this application. It serves as a connector to all the other screens.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.5 Home Screen" + }, + { + "attribute": "page", + "value": "46" + }, + { + "attribute": "figure", + "value": "fig.4.12" + }, + { + "attribute": "source", + "value": "Document" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Home Screen displays activities and reports with visual statistics such as a bar chart.", + "content": "Users can see activities and reports with visual statistical data such as the bar chart as seen in fig.4.12.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.5 Home Screen" + }, + { + "attribute": "page", + "value": "46" + }, + { + "attribute": "figure", + "value": "fig.4.12" + }, + { + "attribute": "topic", + "value": "data visualization" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Home Screen shows recent activities with time and percentage.", + "content": "It also displays recent activities reported with the time and percentage.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.5 Home Screen" + }, + { + "attribute": "page", + "value": "46" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Home Screen is shown only to authenticated users who have passed login validation.", + "content": "This page is only shown to authenticated users, i.e., those who have passed the login validation.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.5 Home Screen" + }, + { + "attribute": "page", + "value": "46" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Create Activity Screen is an interface for tracking activities (referenced by fig.4.13 and fig.4.14).", + "content": "4.4.6 Create Activity Screen This serves as an interface for tracking activities as shown in fig.4.13 and fig.4.14.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.6 Create Activity Screen" + }, + { + "attribute": "page", + "value": "46" + }, + { + "attribute": "figure", + "value": "fig.4.13, fig.4.14" + }, + { + "attribute": "source", + "value": "Document" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "In Create Activity Screen, user selects the activity to track.", + "content": "A user selects the activity to track,", + "attributes": [ + { + "attribute": "section", + "value": "4.4.6 Create Activity Screen" + }, + { + "attribute": "page", + "value": "46" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Create Activity Screen allows setting a time duration, which is optional.", + "content": "set a time duration which is optional", + "attributes": [ + { + "attribute": "section", + "value": "4.4.6 Create Activity Screen" + }, + { + "attribute": "page", + "value": "46" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Create Activity Screen allows setting a report channel.", + "content": "a report channel.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.6 Create Activity Screen" + }, + { + "attribute": "page", + "value": "46" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Time duration on Create Activity Screen constrains tracking to a specific time window.", + "content": "The time duration makes it that activities are only tracked within a certain picked time.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.6 Create Activity Screen" + }, + { + "attribute": "page", + "value": "46" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "The Home Screen is the main interface of the application, acts as a connector to all other screens, displays activities and reports with visual data (bar chart), shows recent activities with time and percentage, and is accessible only to authenticated users.", + "content": "4.4.5 Home Screen This serves as the main interface for this application. It serves as a connector to all the other screens. Users can see activities and reports with visual statistical data such as the bar chart as seen in fig.4.12. It also displays recent activities reported with the time and percentage. This page is only shown to authenticated users, i.e., those who have passed the login validation.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.5" + }, + { + "attribute": "page", + "value": "46-47" + }, + { + "attribute": "figure", + "value": "fig.4.12" + }, + { + "attribute": "topic", + "value": "Home Screen" + }, + { + "attribute": "authentication", + "value": "required" + } + ] + }, + { + "type": "fact", + "insight": "The Create Activity Screen provides an interface for tracking activities; the user can select an activity to track, optionally set a time duration, and choose a report channel; the time duration constrains tracking to a specific time window.", + "content": "4.4.6 Create Activity Screen This serves as an interface for tracking activities as shown in fig.4.13 and fig.4.14. A user selects the activity to track, set a time duration which is optional and a report channel. The time duration makes it that activities are only tracked within a certain picked time.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.6" + }, + { + "attribute": "page", + "value": "46-47" + }, + { + "attribute": "figures", + "value": "fig.4.13 and fig.4.14" + }, + { + "attribute": "input_options", + "value": "activity to track; time duration (optional); report channel" + } + ] + }, + { + "type": "fact", + "insight": "Create Activity Screen allows the user to select an activity to track (UI input for choosing an activity).", + "content": "A user selects the activity to track", + "attributes": [ + { + "attribute": "section", + "value": "4.4.6 Create Activity Screen" + }, + { + "attribute": "page", + "value": "47" + }, + { + "attribute": "figure_referenced", + "value": "fig.4.13, fig.4.14" + }, + { + "attribute": "confidence", + "value": "0.9" + } + ] + }, + { + "type": "fact", + "insight": "Create Activity Screen includes an optional time duration for tracking.", + "content": "set a time duration which is optional", + "attributes": [ + { + "attribute": "section", + "value": "4.4.6 Create Activity Screen" + }, + { + "attribute": "page", + "value": "47" + }, + { + "attribute": "attribute", + "value": "time duration" + } + ] + }, + { + "type": "fact", + "insight": "Create Activity Screen includes a report channel selection.", + "content": "and a report channel.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.6 Create Activity Screen" + }, + { + "attribute": "page", + "value": "47" + } + ] + }, + { + "type": "fact", + "insight": "The time duration defines a time window during which activities are tracked.", + "content": "The time duration makes it that activities are only tracked within a certain picked time.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.6 Create Activity Screen" + }, + { + "attribute": "page", + "value": "47" + } + ] + }, + { + "type": "fact", + "insight": "There are two UI states for Create Activity: Inactive and Active, shown as Figure 4.13 and Figure 4.14.", + "content": "Create Activity Inactive state; Create Activity Active state", + "attributes": [ + { + "attribute": "section", + "value": "4.4.6 Create Activity Screen" + }, + { + "attribute": "page", + "value": "48-49" + }, + { + "attribute": "figure_referenced", + "value": "fig.4.13, fig.4.14" + } + ] + }, + { + "type": "fact", + "insight": "The Activities Screen lists all the activities being tracked by a user.", + "content": "4.4.7 Activities Screen ... lists all the activities that are being tracked by a user.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.7 Activities Screen" + }, + { + "attribute": "page", + "value": "50" + }, + { + "attribute": "figure_reference", + "value": "Figure 4.15" + } + ] + }, + { + "type": "comment", + "insight": "The document uses the terms 'Inactive' and 'Active' states to describe the Create Activity UI, indicating a stateful design.", + "content": "Create Activity Inactive state/Active state", + "attributes": [ + { + "attribute": "section", + "value": "4.4.6 Create Activity Screen" + }, + { + "attribute": "page", + "value": "48-49" + } + ] + }, + { + "type": "comment", + "insight": "The Activities Screen lists all the activities that are being tracked by a user.", + "content": "4.4.7Activities Screen This screen as shown in figure 4.15 lists all the activities that are being tracked by a user.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.7 Activities Screen" + }, + { + "attribute": "page", + "value": "50" + } + ] + }, + { + "type": "comment", + "insight": "The Activity Screen displays information about a specific activity, including its name, how many times it has been reported, and a 30-second video playback of the captured activity.", + "content": "4.4.8 Activity Screen This displays information about a specific activity that is being tracked. It shows the name of the activity, how many times it has been reported and 30 seconds video playback of the captured activity.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.8 Activity Screen" + }, + { + "attribute": "page", + "value": "50-51" + } + ] + }, + { + "type": "fact", + "insight": "Precise findings from the hardware prototype were acquired using SSD to locate bounding boxes, and the OpenCV library was used to draw them.", + "content": "4.5 Summary of Result and Discussion Precise findings from the hardware prototype that were provided to the application for monitoring and tracking were acquired with a view to considering actions. Using Single Shot Detection (SSD) to locate the bounding boxes, the OpenCV library was used to draw them.", + "attributes": [ + { + "attribute": "section", + "value": "4.5 Summary of Result and Discussion" + }, + { + "attribute": "page", + "value": "52" + } + ] + }, + { + "type": "fact", + "insight": "Additionally, the system was able to detect, identify, and analyze a person’s position in order to forecast postural activity.", + "content": "It was also possible to detect, identify, and analyze a person’s position in order to accurately forecast postural activity.", + "attributes": [ + { + "attribute": "section", + "value": "4.5 Summary of Result and Discussion" + }, + { + "attribute": "page", + "value": "52" + } + ] + }, + { + "type": "fact", + "insight": "The conclusion states that the AI-enabled, multi-function activity monitoring and reporting system was a success.", + "content": "5.0 CONCLUSION The design and development of the A.I enabled multi-function-based activity monitoring and reporting system was a success.", + "attributes": [ + { + "attribute": "section", + "value": "5.0 Conclusion" + }, + { + "attribute": "page", + "value": "53" + } + ] + }, + { + "type": "fact", + "insight": "The project used OpenCV DNN and Firebase tooling for its implementation.", + "content": "The system was developed using OpenCV Deep Neural Network (DNN) module and Firebase tooling.", + "attributes": [ + { + "attribute": "section", + "value": "5.0 Conclusion" + }, + { + "attribute": "page", + "value": "53" + } + ] + }, + { + "type": "fact", + "insight": "Video input from a Raspberry Pi camera triggers activity detection and recognition.", + "content": "The video feed obtained from the Pi Camera triggers the process of detecting and recognition of activities.", + "attributes": [ + { + "attribute": "section", + "value": "5.0 Conclusion" + }, + { + "attribute": "page", + "value": "53" + } + ] + }, + { + "type": "fact", + "insight": "The controller continuously communicates with the Firebase Realtime Database to ensure up-to-date activity tracking.", + "content": "The controller constantly communicates with the realtime database to check for up-to date activities to be tracked.", + "attributes": [ + { + "attribute": "section", + "value": "5.0 Conclusion" + }, + { + "attribute": "page", + "value": "53" + } + ] + }, + { + "type": "fact", + "insight": "Motion detection is performed with MobileNetSSD and posture recognition with PoseNet.", + "content": "It performs motion detection using MobileNetSSD and Posture recognition using Posenet.", + "attributes": [ + { + "attribute": "section", + "value": "5.0 Conclusion" + }, + { + "attribute": "page", + "value": "53" + } + ] + }, + { + "type": "fact", + "insight": "Recognized activities are sent to Firebase Realtime Database for tracking and to cloud storage for storage.", + "content": "Recognized activities are then sent over to firebase realtime database for tracking and cloud storage for storage.", + "attributes": [ + { + "attribute": "section", + "value": "5.0 Conclusion" + }, + { + "attribute": "page", + "value": "53" + } + ] + }, + { + "type": "fact", + "insight": "The system can monitor and report multiple activities without human intervention, reducing the need for deploying staff at hospitals or security gates.", + "content": "The system as design was able to monitor and report multiple activities without human intervention, therefore reduces the stress of deploying humans at hospitals or security gates at homes.", + "attributes": [ + { + "attribute": "section", + "value": "5.0 Conclusion" + }, + { + "attribute": "page", + "value": "53" + } + ] + }, + { + "type": "opinion", + "insight": "Recommendations for future work include improving lighting robustness, night vision, live mobile access to video feeds, and camera switching via mobile app.", + "content": "5.2 RECOMMENDATION The designed system met its stated objectives and uncovered the following are some areas for improvement in future work: i. Build support for bad lightening or low camera quality ii. The inclusion of night vision camera to allow the activity recognition model to detect and recognize activities even at night. iii. Live video feeds from the camera should be accessible via the mobile application iv. Camera switching should be made possible from the mobile application.", + "attributes": [ + { + "attribute": "section", + "value": "5.2 Recommendations" + }, + { + "attribute": "page", + "value": "53-54" + } + ] + }, + { + "type": "fact", + "insight": "The 4.4.8 Activity Screen displays information about a specific activity that is being tracked.", + "content": "4.4.8 Activity Screen This displays information about a specific activity that is being tracked. It shows the name of the activity, how many times it has been reported and 30 seconds video playback of the captured activity.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.8 Activity Screen" + }, + { + "attribute": "page", + "value": "50" + } + ] + }, + { + "type": "fact", + "insight": "The Activity Screen shows the name of the activity.", + "content": "It shows the name of the activity,", + "attributes": [ + { + "attribute": "section", + "value": "4.4.8 Activity Screen" + }, + { + "attribute": "page", + "value": "50" + } + ] + }, + { + "type": "fact", + "insight": "The Activity Screen shows how many times the activity has been reported.", + "content": "how many times it has been reported", + "attributes": [ + { + "attribute": "section", + "value": "4.4.8 Activity Screen" + }, + { + "attribute": "page", + "value": "50" + } + ] + }, + { + "type": "fact", + "insight": "The Activity Screen includes 30 seconds of video playback of the captured activity.", + "content": "30 seconds video playback of the captured activity.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.8 Activity Screen" + }, + { + "attribute": "page", + "value": "50" + } + ] + }, + { + "type": "fact", + "insight": "The document states that the interface is shown below in Figure 4.16.", + "content": "The interface is shown below in fig.4.16.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.8 Activity Screen" + }, + { + "attribute": "page", + "value": "50" + } + ] + }, + { + "type": "comment", + "insight": "Figure 4.15 caption indicates the scene is the Activities Screen, and the page references a following figure (Figure 4.16) to illustrate the interface.", + "content": "Figure 4.15: Activities Screen 4.4.8 Activity Screen This displays information about a specific activity that is being tracked. It shows the name of the activity, how many times it has been reported and 30 seconds video playback of the captured activity. The interface is shown below in fig.4.16.", + "attributes": [ + { + "attribute": "section", + "value": "4.4.8 Activity Screen" + }, + { + "attribute": "page", + "value": "50" + } + ] + }, + { + "type": "comment", + "insight": "No concrete data values (e.g., activity name or report count) are provided on this page.", + "content": "It shows the name of the activity, how many times it has been reported", + "attributes": [ + { + "attribute": "section", + "value": "4.4.8 Activity Screen" + }, + { + "attribute": "page", + "value": "50" + } + ] + }, + { + "type": "fact", + "insight": "SSD is used to locate the bounding boxes and OpenCV is used to draw them.", + "content": "Using Single Shot Detection (SSD) to locate the bounding boxes, the OpenCV library was used to draw them.", + "attributes": [ + { + "attribute": "section", + "value": "4.5 Summary of Result and Discussion" + }, + { + "attribute": "source_page", + "value": "51" + }, + { + "attribute": "technologies", + "value": "SSD, OpenCV" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "It was possible to detect, identify, and analyze a person\\'s position in order to forecast postural activity.", + "content": "It was also possible to detect, identify, and analyze a person\\'s position in order to accurately forecast postural activity.", + "attributes": [ + { + "attribute": "section", + "value": "4.5 Summary of Result and Discussion" + }, + { + "attribute": "source_page", + "value": "51" + }, + { + "attribute": "topic", + "value": "postural activity forecasting" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "AI-enabled multi-function-based activity monitoring and reporting system design was a success.", + "content": "The design and development of the A.I enabled multi-function-based activity monitoring and reporting system was a success.", + "attributes": [ + { + "attribute": "section", + "value": "5.0 Conclusion" + }, + { + "attribute": "source_page", + "value": "52" + }, + { + "attribute": "tone", + "value": "positive" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "The system was developed using OpenCV Deep Neural Network (DNN) module and Firebase tooling.", + "content": "The system was developed using OpenCV Deep Neural Network (DNN) module and Firebase tooling.", + "attributes": [ + { + "attribute": "section", + "value": "5.0 Conclusion" + }, + { + "attribute": "source_page", + "value": "52" + }, + { + "attribute": "technologies", + "value": "OpenCV DNN, Firebase" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Video feed from the Pi Camera triggers the process of detecting and recognition of activities.", + "content": "The video feed obtained from the Pi Camera triggers the process of detecting and recognition of activities.", + "attributes": [ + { + "attribute": "section", + "value": "5.0 Conclusion" + }, + { + "attribute": "source_page", + "value": "52" + }, + { + "attribute": "camera", + "value": "Pi Camera" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "The controller communicates with the realtime database to check for up-to-date activities to be tracked.", + "content": "The controller constantly communicates with the realtime database to check for up-to-date activities to be tracked.", + "attributes": [ + { + "attribute": "section", + "value": "5.0 Conclusion" + }, + { + "attribute": "source_page", + "value": "52" + }, + { + "attribute": "component", + "value": "controller" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Motion detection is performed using MobileNetSSD and posture recognition using PoseNet.", + "content": "It performs motion detection using MobileNetSSD and Posture recognition using Posenet.", + "attributes": [ + { + "attribute": "section", + "value": "5.0 Conclusion" + }, + { + "attribute": "source_page", + "value": "52" + }, + { + "attribute": "models", + "value": "MobileNetSSD, PoseNet" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Recognized activities are sent to the Firebase realtime database for tracking and to cloud storage for storage.", + "content": "Recognized activities are then sent over to firebase realtime database for tracking and cloud storage for storage.", + "attributes": [ + { + "attribute": "section", + "value": "5.0 Conclusion" + }, + { + "attribute": "source_page", + "value": "52" + }, + { + "attribute": "storage", + "value": "Firebase realtime database, cloud storage" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "The system can monitor and report multiple activities without human intervention, reducing the need for deploying humans at hospitals or security gates at homes.", + "content": "The system as design was able to monitor and report multiple activities without human intervention, therefore reduces the stress of deploying humans at hospitals or security gates at homes.", + "attributes": [ + { + "attribute": "section", + "value": "5.0 Conclusion" + }, + { + "attribute": "source_page", + "value": "52" + }, + { + "attribute": "impact", + "value": "reduces human deployment at hospitals/security gates" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "opinion", + "insight": "Build support for bad lightening or low camera quality.", + "content": "i. Build support for bad lightening or low camera quality", + "attributes": [ + { + "attribute": "section", + "value": "5.2 Recommendations" + }, + { + "attribute": "source_page", + "value": "52" + }, + { + "attribute": "topic", + "value": "lighting conditions" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "opinion", + "insight": "The inclusion of night vision camera to allow the activity recognition model to detect and recognize activities even at night.", + "content": "ii. The inclusion of night vision camera to allow the activity recognition model to detect and recognize activities even at night.", + "attributes": [ + { + "attribute": "section", + "value": "5.2 Recommendations" + }, + { + "attribute": "source_page", + "value": "52" + }, + { + "attribute": "topic", + "value": "night vision" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "opinion", + "insight": "Live video feeds from the camera should be accessible via the mobile application.", + "content": "iii. Live video feeds from the camera should be accessible via the mobile application", + "attributes": [ + { + "attribute": "section", + "value": "5.2 Recommendations" + }, + { + "attribute": "source_page", + "value": "52" + }, + { + "attribute": "topic", + "value": "mobile app access to feeds" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "opinion", + "insight": "Camera switching should be made possible from the mobile application.", + "content": "iv. Camera switching should be made possible from the mobile application.", + "attributes": [ + { + "attribute": "section", + "value": "5.2 Recommendations" + }, + { + "attribute": "source_page", + "value": "52" + }, + { + "attribute": "topic", + "value": "camera switching" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "The system was developed using OpenCV Deep Neural Network (DNN) module and Firebase tooling.", + "content": "The system was developed using OpenCV Deep Neural Network (DNN) module and Firebase tooling.", + "attributes": [ + { + "attribute": "section", + "value": "5.1" + }, + { + "attribute": "source", + "value": "Chapter 5" + } + ] + }, + { + "type": "fact", + "insight": "The system was designed and developed in accordance with the objectives in Section 1.3, the outlined methodologies in Section 3.2, the overall design and flow explained in Section 3.3.", + "content": "The system was designed and developed in accordance with the objectives in Section 1.3, the outlined methodologies in Section 3.2, the overall design and flow explained in Section 3.3.", + "attributes": [ + { + "attribute": "section", + "value": "5.1" + }, + { + "attribute": "source", + "value": "Chapter 5" + } + ] + }, + { + "type": "fact", + "insight": "The video feed obtained from the Pi Camera triggers the process of detecting and recognition of activities.", + "content": "The video feed obtained from the Pi Camera triggers the process of detecting and recognition of activities.", + "attributes": [ + { + "attribute": "section", + "value": "5.1" + }, + { + "attribute": "source", + "value": "Chapter 5" + } + ] + }, + { + "type": "fact", + "insight": "The controller constantly communicates with the realtime database to check for up-to-date activities to be tracked.", + "content": "The controller constantly communicates with the realtime database to check for up-to-date activities to be tracked.", + "attributes": [ + { + "attribute": "section", + "value": "5.1" + }, + { + "attribute": "source", + "value": "Chapter 5" + } + ] + }, + { + "type": "fact", + "insight": "It performs motion detection using MobileNetSSD and Posture recognition using Posenet.", + "content": "It performs motion detection using MobileNetSSD and Posture recognition using Posenet.", + "attributes": [ + { + "attribute": "section", + "value": "5.1" + }, + { + "attribute": "source", + "value": "Chapter 5" + } + ] + }, + { + "type": "fact", + "insight": "Recognized activities are then sent over to firebase realtime database for tracking and cloud storage for storage.", + "content": "Recognized activities are then sent over to firebase realtime database for tracking and cloud storage for storage.", + "attributes": [ + { + "attribute": "section", + "value": "5.1" + }, + { + "attribute": "source", + "value": "Chapter 5" + } + ] + }, + { + "type": "fact", + "insight": "The system as design was able to monitor and report multiple activities without human intervention, therefore reduces the stress of deploying humans at hospitals or security gates at homes.", + "content": "The system as design was able to monitor and report multiple activities without human intervention, therefore reduces the stress of deploying humans at hospitals or security gates at homes.", + "attributes": [ + { + "attribute": "section", + "value": "5.1" + }, + { + "attribute": "source", + "value": "Chapter 5" + } + ] + }, + { + "type": "opinion", + "insight": "i. Build support for bad lightening or low camera quality", + "content": "i. Build support for bad lightening or low camera quality", + "attributes": [ + { + "attribute": "section", + "value": "5.2" + }, + { + "attribute": "source", + "value": "Chapter 5" + } + ] + }, + { + "type": "opinion", + "insight": "ii. The inclusion of night vision camera to allow the activity recognition model to detect and recognize activities even at night.", + "content": "ii. The inclusion of night vision camera to allow the activity recognition model to detect and recognize activities even at night.", + "attributes": [ + { + "attribute": "section", + "value": "5.2" + }, + { + "attribute": "source", + "value": "Chapter 5" + } + ] + }, + { + "type": "opinion", + "insight": "iii. Live video feeds from the camera should be accessible via the mobile application", + "content": "iii. Live video feeds from the camera should be accessible via the mobile application", + "attributes": [ + { + "attribute": "section", + "value": "5.2" + }, + { + "attribute": "source", + "value": "Chapter 5" + } + ] + }, + { + "type": "opinion", + "insight": "iv. Camera switching should be made possible from the mobile application.", + "content": "iv. Camera switching should be made possible from the mobile application.", + "attributes": [ + { + "attribute": "section", + "value": "5.2" + }, + { + "attribute": "source", + "value": "Chapter 5" + } + ] + }, + { + "type": "fact", + "insight": "Vision-based monitoring reference: Chaaraoui et al. 2014 cite A vision-based system for intelligent monitoring: Human behaviour analysis and privacy by context (Sensors 2014; 14:8895-8925).", + "content": "Chaaraoui, A. A., Padilla-López, J. R., Ferrández-Pastor, F. J., Nieto-Hidalgo, M., & Flórez-Revuelta, F. (2014). A vision-based system for intelligent monitoring: Human behaviour analysis and privacy by context. A Vision-Based System for Intelligent Monitoring: Human Behaviour Analysis and Privacy by Context Alexandros, 14(5), 8895–8925. https://doi.org/10.3390/s140508895", + "attributes": [ + { + "attribute": "source", + "value": "Sensors (2014) 14:8895-8925" + }, + { + "attribute": "authors", + "value": "Chaaraoui A.A.; Padilla-López J.R.; Ferrández-Pastor F.J.; Nieto-Hidalgo M.; Flórez-Revuelta F." + }, + { + "attribute": "year", + "value": "2014" + }, + { + "attribute": "title", + "value": "A vision-based system for intelligent monitoring: Human behaviour analysis and privacy by context" + }, + { + "attribute": "venue", + "value": "Sensors" + }, + { + "attribute": "pages", + "value": "8895-8925" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.3390/s140508895" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Imaging database foundational: Deng et al. 2009 Imagenet — large-scale hierarchical image database (CVPR 2009).", + "content": "Deng, J., Dong, W., Socher, R., Li, L. J., Li, K., & Fei-Fei, L. (2009). Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition (pp. 248-255).", + "attributes": [ + { + "attribute": "source", + "value": "IEEE CVPR 2009" + }, + { + "attribute": "authors", + "value": "Deng J.; Dong W.; Socher R.; Li L.J.; Li K.; Fei-Fei L." + }, + { + "attribute": "year", + "value": "2009" + }, + { + "attribute": "title", + "value": "Imagenet: A large-scale hierarchical image database" + }, + { + "attribute": "venue", + "value": "IEEE CVPR 2009" + }, + { + "attribute": "pages", + "value": "248-255" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Deep learning fundamentals: Mishra & Gupta 2017 Deep Machine Learning and Neural Networks: An Overview (IJ-AI).", + "content": "Mishra, C., & Gupta, D. L. (2017). Deep Machine Learning and Neural Networks: An Overview. IAES International Journal of Artificial Intelligence (IJ-AI), 6(2), 66-73. https://doi.org/10.11591/ijai.v6.i2.pp66-73", + "attributes": [ + { + "attribute": "source", + "value": "IJ-AI (2017) 6(2) 66-73" + }, + { + "attribute": "authors", + "value": "Mishra C.; Gupta D.L." + }, + { + "attribute": "year", + "value": "2017" + }, + { + "attribute": "title", + "value": "Deep Machine Learning and Neural Networks: An Overview" + }, + { + "attribute": "venue", + "value": "IAES IJ-AI" + }, + { + "attribute": "pages", + "value": "66-73" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.11591/ijai.v6.i2.pp66-73" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Activity monitoring for elderly: Tee et al. 2015 An activity monitoring system for elderly (ARPN Journal).", + "content": "Tee, K. S., Zulkifli, A. H. B., & Soon, C. F. (2015). An activity monitoring system for elderly. ARPN Journal of Engineering and Applied Sciences, 10(18), 8467–8472.", + "attributes": [ + { + "attribute": "source", + "value": "ARPN Journal of Engineering and Applied Sciences (2015) 10(18) 8467-8472" + }, + { + "attribute": "authors", + "value": "Tee K.S.; Zulkifli A.H.B.; Soon C.F." + }, + { + "attribute": "year", + "value": "2015" + }, + { + "attribute": "title", + "value": "An activity monitoring system for elderly" + }, + { + "attribute": "venue", + "value": "ARPN Journal of Engineering and Applied Sciences" + }, + { + "attribute": "pages", + "value": "8467-8472" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Defining AI: Wang 2019 On Defining Artificial Intelligence (JAGI).", + "content": "Wang, P. (2019). On Defining Artificial Intelligence. Journal of Artificial General Intelligence , 10 (2), 1–37. https://doi.org/10.2478/jagi-2019-0002", + "attributes": [ + { + "attribute": "source", + "value": "Journal of Artificial General Intelligence (2019) 10(2) 1-37" + }, + { + "attribute": "author", + "value": "Wang P." + }, + { + "attribute": "year", + "value": "2019" + }, + { + "attribute": "title", + "value": "On Defining Artificial Intelligence" + }, + { + "attribute": "venue", + "value": "Journal of Artificial General Intelligence" + }, + { + "attribute": "pages", + "value": "1-37" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.2478/jagi-2019-0002" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Competitive advantage via machine learning: Attaran & Deb 2018 (IJKEDM)", + "content": "Attaran, M., & Deb, P. (2018). Machine Learning : The New ’ Big Thing ’ for Competitive Advantage International Journal of Knowledge Engineering and Data Mining, 5(4), 277-305. https://doi.org/10.1504/IJKEDM.2018.10015621.", + "attributes": [ + { + "attribute": "source", + "value": "IJKEDM (2018) 5(4) 277-305" + }, + { + "attribute": "authors", + "value": "Attaran M.; Deb P." + }, + { + "attribute": "year", + "value": "2018" + }, + { + "attribute": "title", + "value": "Machine Learning: The New 'Big Thing' for Competitive Advantage" + }, + { + "attribute": "venue", + "value": "International Journal of Knowledge Engineering and Data Mining (IJKEDM)" + }, + { + "attribute": "pages", + "value": "277-305" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.1504/IJKEDM.2018.10015621" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Fast R-CNN: Girshick 2015 (CVPR)", + "content": "Girshick, R. (2015). Fast r-cnn. In Proceedings of the IEEE international conference on computer vision (pp. 1440-1448).", + "attributes": [ + { + "attribute": "source", + "value": "CVPR 2015" + }, + { + "attribute": "author", + "value": "Girshick R." + }, + { + "attribute": "year", + "value": "2015" + }, + { + "attribute": "title", + "value": "Fast R-CNN" + }, + { + "attribute": "venue", + "value": "IEEE CVPR 2015" + }, + { + "attribute": "pages", + "value": "1440-1448" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Feature Pyramid Networks for detection: Lin et al. 2017 (CVPR)", + "content": "Lin, T. Y., Dollár, P., Girshick, R., He, K., Hariharan, B., & Belongie, S. (2017). Feature pyramid networks for object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 2117-2125).", + "attributes": [ + { + "attribute": "source", + "value": "CVPR 2017" + }, + { + "attribute": "authors", + "value": "Lin T.Y.; Dollár P.; Girshick R.; He K.; Hariharan B.; Belongie S." + }, + { + "attribute": "year", + "value": "2017" + }, + { + "attribute": "title", + "value": "Feature pyramid networks for object detection" + }, + { + "attribute": "venue", + "value": "CVPR 2017" + }, + { + "attribute": "pages", + "value": "2117-2125" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "AI in disasters: Lu et al. 2021 (Disaster Medicine and Public Health Preparedness)", + "content": "Lu, S., Christie, G. A., Nguyen, T. T., Freeman, J. D, and Hsu, E. B., (2021). Applications of Artificial Intelligence and Machine Learning in Disasters and Public Health Emergencies. Disaster Medicine and Public Health Preparedness, 1-8.", + "attributes": [ + { + "attribute": "source", + "value": "Disaster Medicine and Public Health Preparedness (2021) 1-8" + }, + { + "attribute": "authors", + "value": "Lu S.; Christie G.A.; Nguyen T.T.; Freeman J.D.; Hsu E.B." + }, + { + "attribute": "year", + "value": "2021" + }, + { + "attribute": "title", + "value": "Applications of Artificial Intelligence and Machine Learning in Disasters and Public Health Emergencies" + }, + { + "attribute": "venue", + "value": "Disaster Medicine and Public Health Preparedness" + }, + { + "attribute": "pages", + "value": "1-8" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "YOLOv2 for vehicle tracking: Malaainine et al. 2021 (JGIS)", + "content": "Malaainine, M. E, I., Lechgar, H., & Rhinane, H. (2021). Y OLOv2 Deep Learning Model and GIS Based Algorithms for Vehicle Tracking. Journal of Geographic Information System, 2021, 13, 395-409. https://doi.org/10.4236/jgis.2021.134022.", + "attributes": [ + { + "attribute": "source", + "value": "Journal of Geographic Information System (2021) 13:395-409" + }, + { + "attribute": "authors", + "value": "Malaainine M.E.I.; Lechgar H.; Rhinane H." + }, + { + "attribute": "year", + "value": "2021" + }, + { + "attribute": "title", + "value": "YOLOv2 Deep Learning Model and GIS Based Algorithms for Vehicle Tracking" + }, + { + "attribute": "venue", + "value": "Journal of Geographic Information System" + }, + { + "attribute": "pages", + "value": "395-409" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.4236/jgis.2021.134022" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Crisis tweet detection: Kruspe et al. 2021 (NHESS)", + "content": "Kruspe, A., Kersten, J., & Klan, F. (2021). Review article: Detection of actionable tweets in crisis events. Natural Hazards Earth System Science., 21, 1825–1845. https://doi.org/10.5194/nhess-21-1825-2021.", + "attributes": [ + { + "attribute": "source", + "value": "NHESS (2021) 21:1825-1845" + }, + { + "attribute": "authors", + "value": "Kruspe A.; Kersten J.; Klan F." + }, + { + "attribute": "year", + "value": "2021" + }, + { + "attribute": "title", + "value": "Detection of actionable tweets in crisis events" + }, + { + "attribute": "venue", + "value": "Natural Hazards Earth System Science" + }, + { + "attribute": "pages", + "value": "1825-1845" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.5194/nhess-21-1825-2021" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Deeper CNNs: Szegedy et al. 2015 Going Deeper with Convolutions (CVPR).", + "content": "Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., & Rabinovich, A. (2015). Going deeper with convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 1-9).", + "attributes": [ + { + "attribute": "source", + "value": "CVPR 2015" + }, + { + "attribute": "authors", + "value": "Szegedy C.; Liu W.; Jia Y.; Sermanet P.; Reed S.; Anguelov D.; Rabinovich A." + }, + { + "attribute": "year", + "value": "2015" + }, + { + "attribute": "title", + "value": "Going deeper with convolutions" + }, + { + "attribute": "venue", + "value": "CVPR 2015" + }, + { + "attribute": "pages", + "value": "1-9" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "R-CNN family: Fast R-CNN (Girshick 2015)", + "content": "Girshick, R. (2015). Fast r-cnn. In Proceedings of the IEEE international conference on computer vision (pp. 1440-1448).", + "attributes": [ + { + "attribute": "source", + "value": "CVPR 2015" + }, + { + "attribute": "authors", + "value": "Girshick R." + }, + { + "attribute": "year", + "value": "2015" + }, + { + "attribute": "title", + "value": "Fast R-CNN" + }, + { + "attribute": "venue", + "value": "IEEE CVPR 2015" + }, + { + "attribute": "pages", + "value": "1440-1448" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Object detection with FPN: Lin et al. 2017 (CVPR)", + "content": "Lin, T. Y., Dollár, P., Girshick, R., He, K., Hariharan, B., & Belongie, S. (2017). Feature pyramid networks for object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 2117-2125).", + "attributes": [ + { + "attribute": "source", + "value": "CVPR 2017" + }, + { + "attribute": "authors", + "value": "Lin T.Y.; Dollár P.; Girshick R.; He K.; Hariharan B.; Belongie S." + }, + { + "attribute": "year", + "value": "2017" + }, + { + "attribute": "title", + "value": "Feature pyramid networks for object detection" + }, + { + "attribute": "venue", + "value": "CVPR 2017" + }, + { + "attribute": "pages", + "value": "2117-2125" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "AI in disasters: Lu et al. 2021 applications in disasters and public health emergencies", + "content": "Lu, S., Christie, G. A., Nguyen, T. T., Freeman, J. D, and Hsu, E. B., (2021). Applications of Artificial Intelligence and Machine Learning in Disasters and Public Health Emergencies. Disaster Medicine and Public Health Preparedness, 1-8.", + "attributes": [ + { + "attribute": "source", + "value": "Disaster Medicine and Public Health Preparedness (2021) 1-8" + }, + { + "attribute": "authors", + "value": "Lu S.; Christie G.A.; Nguyen T.T.; Freeman J.D.; Hsu E.B." + }, + { + "attribute": "year", + "value": "2021" + }, + { + "attribute": "title", + "value": "Applications of Artificial Intelligence and Machine Learning in Disasters and Public Health Emergencies" + }, + { + "attribute": "venue", + "value": "Disaster Medicine and Public Health Preparedness" + }, + { + "attribute": "pages", + "value": "1-8" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "YOLOv2 for vehicle tracking: Malaainine et al. 2021 (JOIS/JGIS)", + "content": "Malaainine, M. E, I., Lechgar, H., & Rhinane, H. (2021). Y OLOv2 Deep Learning Model and GIS Based Algorithms for Vehicle Tracking. Journal of Geographic Information System, 2021, 13, 395-409. https://doi.org/10.4236/jgis.2021.134022.", + "attributes": [ + { + "attribute": "source", + "value": "Journal of Geographic Information System (2021) 13:395-409" + }, + { + "attribute": "authors", + "value": "Malaainine M.E.I.; Lechgar H.; Rhinane H." + }, + { + "attribute": "year", + "value": "2021" + }, + { + "attribute": "title", + "value": "YOLOv2 Deep Learning Model and GIS Based Algorithms for Vehicle Tracking" + }, + { + "attribute": "venue", + "value": "Journal of Geographic Information System" + }, + { + "attribute": "pages", + "value": "395-409" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.4236/jgis.2021.134022" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Crisis tweets detection: Kruspe et al. 2021 NHESS review article", + "content": "Kruspe, A., Kersten, J., & Klan, F. (2021). Review article: Detection of actionable tweets in crisis events. Natural Hazards Earth System Science., 21, 1825–1845. https://doi.org/10.5194/nhess-21-1825-2021.", + "attributes": [ + { + "attribute": "source", + "value": "NHESS (2021) 21:1825-1845" + }, + { + "attribute": "authors", + "value": "Kruspe A.; Kersten J.; Klan F." + }, + { + "attribute": "year", + "value": "2021" + }, + { + "attribute": "title", + "value": "Detection of actionable tweets in crisis events" + }, + { + "attribute": "venue", + "value": "Natural Hazards Earth System Science" + }, + { + "attribute": "pages", + "value": "1825-1845" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.5194/nhess-21-1825-2021" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Deeper CNNs: Szegedy et al. 2015 Going Deeper with Convolutions (CVPR)", + "content": "Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., & Rabinovich, A. (2015). Going deeper with convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 1-9).", + "attributes": [ + { + "attribute": "source", + "value": "CVPR 2015" + }, + { + "attribute": "authors", + "value": "Szegedy C.; Liu W.; Jia Y.; Sermanet P.; Reed S.; Anguelov D.; Rabinovich A." + }, + { + "attribute": "year", + "value": "2015" + }, + { + "attribute": "title", + "value": "Going deeper with convolutions" + }, + { + "attribute": "venue", + "value": "CVPR 2015" + }, + { + "attribute": "pages", + "value": "1-9" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Faster R-CNN: Girshick 2015 (CVPR)", + "content": "Girshick, R. (2015). Fast r-cnn. In Proceedings of the IEEE international conference on computer vision (pp. 1440-1448).", + "attributes": [ + { + "attribute": "source", + "value": "CVPR 2015" + }, + { + "attribute": "authors", + "value": "Girshick R." + }, + { + "attribute": "year", + "value": "2015" + }, + { + "attribute": "title", + "value": "Fast R-CNN" + }, + { + "attribute": "venue", + "value": "IEEE CVPR 2015" + }, + { + "attribute": "pages", + "value": "1440-1448" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Feature pyramids for detection: Lin et al. 2017 (CVPR)", + "content": "Lin, T. Y., Dollár, P., Girshick, R., He, K., Hariharan, B., & Belongie, S. (2017). Feature pyramid networks for object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 2117-2125).", + "attributes": [ + { + "attribute": "source", + "value": "CVPR 2017" + }, + { + "attribute": "authors", + "value": "Lin T.Y.; Dollár P.; Girshick R.; He K.; Hariharan B.; Belongie S." + }, + { + "attribute": "year", + "value": "2017" + }, + { + "attribute": "title", + "value": "Feature pyramid networks for object detection" + }, + { + "attribute": "venue", + "value": "CVPR 2017" + }, + { + "attribute": "pages", + "value": "2117-2125" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Disaster AI applications: Lu et al. 2021 Disaster Medicine and Public Health Preparedness", + "content": "Lu, S., Christie, G. A., Nguyen, T. T., Freeman, J. D, and Hsu, E. B., (2021). Applications of Artificial Intelligence and Machine Learning in Disasters and Public Health Emergencies. Disaster Medicine and Public Health Preparedness, 1-8.", + "attributes": [ + { + "attribute": "source", + "value": "Disaster Medicine and Public Health Preparedness (2021) 1-8" + }, + { + "attribute": "authors", + "value": "Lu S.; Christie G.A.; Nguyen T.T.; Freeman J.D.; Hsu E.B." + }, + { + "attribute": "year", + "value": "2021" + }, + { + "attribute": "title", + "value": "Applications of Artificial Intelligence and Machine Learning in Disasters and Public Health Emergencies" + }, + { + "attribute": "venue", + "value": "Disaster Medicine and Public Health Preparedness" + }, + { + "attribute": "pages", + "value": "1-8" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "YOLOv2 vehicle tracking: Malaainine et al. 2021 (JGIS)", + "content": "Malaainine, M. E, I., Lechgar, H., & Rhinane, H. (2021). Y OLOv2 Deep Learning Model and GIS Based Algorithms for Vehicle Tracking. Journal of Geographic Information System, 2021, 13, 395-409. https://doi.org/10.4236/jgis.2021.134022.", + "attributes": [ + { + "attribute": "source", + "value": "Journal of Geographic Information System (2021) 13:395-409" + }, + { + "attribute": "authors", + "value": "Malaainine M.E.I.; Lechgar H.; Rhinane H." + }, + { + "attribute": "year", + "value": "2021" + }, + { + "attribute": "title", + "value": "YOLOv2 Deep Learning Model and GIS Based Algorithms for Vehicle Tracking" + }, + { + "attribute": "venue", + "value": "Journal of Geographic Information System" + }, + { + "attribute": "pages", + "value": "395-409" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.4236/jgis.2021.134022" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Crisis tweet detection: Kruspe et al. 2021 NHESS", + "content": "Kruspe, A., Kersten, J., & Klan, F. (2021). Review article: Detection of actionable tweets in crisis events. Natural Hazards Earth System Science., 21, 1825–1845. https://doi.org/10.5194/nhess-21-1825-2021.", + "attributes": [ + { + "attribute": "source", + "value": "NHESS (2021) 21:1825-1845" + }, + { + "attribute": "authors", + "value": "Kruspe A.; Kersten J.; Klan F." + }, + { + "attribute": "year", + "value": "2021" + }, + { + "attribute": "title", + "value": "Detection of actionable tweets in crisis events" + }, + { + "attribute": "venue", + "value": "Natural Hazards Earth System Science" + }, + { + "attribute": "pages", + "value": "1825-1845" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.5194/nhess-21-1825-2021" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "CNN depth: Szegedy 2015 Going Deeper with Convolutions (CVPR)", + "content": "Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., & Rabinovich, A. (2015). Going deeper with convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 1-9).", + "attributes": [ + { + "attribute": "source", + "value": "CVPR 2015" + }, + { + "attribute": "authors", + "value": "Szegedy C.; Liu W.; Jia Y.; Sermanet P.; Reed S.; Anguelov D.; Rabinovich A." + }, + { + "attribute": "year", + "value": "2015" + }, + { + "attribute": "title", + "value": "Going deeper with convolutions" + }, + { + "attribute": "venue", + "value": "CVPR 2015" + }, + { + "attribute": "pages", + "value": "1-9" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Faster R-CNN: Girshick 2015 (CVPR)", + "content": "Girshick, R. (2015). Fast r-cnn. In Proceedings of the IEEE international conference on computer vision (pp. 1440-1448).", + "attributes": [ + { + "attribute": "source", + "value": "CVPR 2015" + }, + { + "attribute": "authors", + "value": "Girshick R." + }, + { + "attribute": "year", + "value": "2015" + }, + { + "attribute": "title", + "value": "Fast R-CNN" + }, + { + "attribute": "venue", + "value": "IEEE CVPR 2015" + }, + { + "attribute": "pages", + "value": "1440-1448" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Feature pyramids for detection: Lin et al. 2017 (CVPR)", + "content": "Lin, T. Y., Dollár, P., Girshick, R., He, K., Hariharan, B., & Belongie, S. (2017). Feature pyramid networks for object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 2117-2125).", + "attributes": [ + { + "attribute": "source", + "value": "CVPR 2017" + }, + { + "attribute": "authors", + "value": "Lin T.Y.; Dollár P.; Girshick R.; He K.; Hariharan B.; Belongie S." + }, + { + "attribute": "year", + "value": "2017" + }, + { + "attribute": "title", + "value": "Feature pyramid networks for object detection" + }, + { + "attribute": "venue", + "value": "CVPR 2017" + }, + { + "attribute": "pages", + "value": "2117-2125" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Disaster AI applications: Lu et al. 2021 DLPMHP", + "content": "Lu, S., Christie, G. A., Nguyen, T. T., Freeman, J. D, and Hsu, E. B., (2021). Applications of Artificial Intelligence and Machine Learning in Disasters and Public Health Emergencies. Disaster Medicine and Public Health Preparedness, 1-8.", + "attributes": [ + { + "attribute": "source", + "value": "Disaster Medicine and Public Health Preparedness (2021) 1-8" + }, + { + "attribute": "authors", + "value": "Lu S.; Christie G.A.; Nguyen T.T.; Freeman J.D.; Hsu E.B." + }, + { + "attribute": "year", + "value": "2021" + }, + { + "attribute": "title", + "value": "Applications of Artificial Intelligence and Machine Learning in Disasters and Public Health Emergencies" + }, + { + "attribute": "venue", + "value": "Disaster Medicine and Public Health Preparedness" + }, + { + "attribute": "pages", + "value": "1-8" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "YOLOv2 for vehicle tracking: Malaainine et al. 2021 (JGIS)", + "content": "Malaainine, M. E, I., Lechgar, H., & Rhinane, H. (2021). Y OLOv2 Deep Learning Model and GIS Based Algorithms for Vehicle Tracking. Journal of Geographic Information System, 2021, 13, 395-409. https://doi.org/10.4236/jgis.2021.134022.", + "attributes": [ + { + "attribute": "source", + "value": "Journal of Geographic Information System (2021) 13:395-409" + }, + { + "attribute": "authors", + "value": "Malaainine M.E.I.; Lechgar H.; Rhinane H." + }, + { + "attribute": "year", + "value": "2021" + }, + { + "attribute": "title", + "value": "YOLOv2 Deep Learning Model and GIS Based Algorithms for Vehicle Tracking" + }, + { + "attribute": "venue", + "value": "Journal of Geographic Information System" + }, + { + "attribute": "pages", + "value": "395-409" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.4236/jgis.2021.134022" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Crisis tweets detection: Kruspe et al. 2021 NHESS review article", + "content": "Kruspe, A., Kersten, J., & Klan, F. (2021). Review article: Detection of actionable tweets in crisis events. Natural Hazards Earth System Science., 21, 1825–1845. https://doi.org/10.5194/nhess-21-1825-2021.", + "attributes": [ + { + "attribute": "source", + "value": "NHESS (2021) 21:1825-1845" + }, + { + "attribute": "authors", + "value": "Kruspe A.; Kersten J.; Klan F." + }, + { + "attribute": "year", + "value": "2021" + }, + { + "attribute": "title", + "value": "Detection of actionable tweets in crisis events" + }, + { + "attribute": "venue", + "value": "Natural Hazards Earth System Science" + }, + { + "attribute": "pages", + "value": "1825-1845" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.5194/nhess-21-1825-2021" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Deeper CNNs: Szegedy et al. 2015 Going Deeper with Convolutions (CVPR)", + "content": "Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., & Rabinovich, A. (2015). Going deeper with convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 1-9).", + "attributes": [ + { + "attribute": "source", + "value": "CVPR 2015" + }, + { + "attribute": "authors", + "value": "Szegedy C.; Liu W.; Jia Y.; Sermanet P.; Reed S.; Anguelov D.; Rabinovich A." + }, + { + "attribute": "year", + "value": "2015" + }, + { + "attribute": "title", + "value": "Going deeper with convolutions" + }, + { + "attribute": "venue", + "value": "CVPR 2015" + }, + { + "attribute": "pages", + "value": "1-9" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Faster R-CNN: Girshick 2015 (CVPR)", + "content": "Girshick, R. (2015). Fast r-cnn. In Proceedings of the IEEE international conference on computer vision (pp. 1440-1448).", + "attributes": [ + { + "attribute": "source", + "value": "CVPR 2015" + }, + { + "attribute": "authors", + "value": "Girshick R." + }, + { + "attribute": "year", + "value": "2015" + }, + { + "attribute": "title", + "value": "Fast R-CNN" + }, + { + "attribute": "venue", + "value": "IEEE CVPR 2015" + }, + { + "attribute": "pages", + "value": "1440-1448" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "FPN for object detection: Lin et al. 2017 CVPR", + "content": "Lin, T. Y., Dollár, P., Girshick, R., He, K., Hariharan, B., & Belongie, S. (2017). Feature pyramid networks for object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 2117-2125).", + "attributes": [ + { + "attribute": "source", + "value": "CVPR 2017" + }, + { + "attribute": "authors", + "value": "Lin T.Y.; Dollár P.; Girshick R.; He K.; Hariharan B.; Belongie S." + }, + { + "attribute": "year", + "value": "2017" + }, + { + "attribute": "title", + "value": "Feature pyramid networks for object detection" + }, + { + "attribute": "venue", + "value": "CVPR 2017" + }, + { + "attribute": "pages", + "value": "2117-2125" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "AI in disasters: Lu et al. 2021 again", + "content": "Lu, S., Christie, G. A., Nguyen, T. T., Freeman, J. D, and Hsu, E. B., (2021). Applications of Artificial Intelligence and Machine Learning in Disasters and Public Health Emergencies. Disaster Medicine and Public Health Preparedness, 1-8.", + "attributes": [ + { + "attribute": "source", + "value": "Disaster Medicine and Public Health Preparedness (2021) 1-8" + }, + { + "attribute": "authors", + "value": "Lu S.; Christie G.A.; Nguyen T.T.; Freeman J.D.; Hsu E.B." + }, + { + "attribute": "year", + "value": "2021" + }, + { + "attribute": "title", + "value": "Applications of Artificial Intelligence and Machine Learning in Disasters and Public Health Emergencies" + }, + { + "attribute": "venue", + "value": "Disaster Medicine and Public Health Preparedness" + }, + { + "attribute": "pages", + "value": "1-8" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "YOLOv2 for vehicle tracking: Malaainine et al. 2021 JGIS", + "content": "Malaainine, M. E, I., Lechgar, H., & Rhinane, H. (2021). Y OLOv2 Deep Learning Model and GIS Based Algorithms for Vehicle Tracking. Journal of Geographic Information System, 2021, 13, 395-409. https://doi.org/10.4236/jgis.2021.134022.", + "attributes": [ + { + "attribute": "source", + "value": "Journal of Geographic Information System (2021) 13:395-409" + }, + { + "attribute": "authors", + "value": "Malaainine M.E.I.; Lechgar H.; Rhinane H." + }, + { + "attribute": "year", + "value": "2021" + }, + { + "attribute": "title", + "value": "YOLOv2 Deep Learning Model and GIS Based Algorithms for Vehicle Tracking" + }, + { + "attribute": "venue", + "value": "Journal of Geographic Information System" + }, + { + "attribute": "pages", + "value": "395-409" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.4236/jgis.2021.134022" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Crisis tweet detection review: Kruspe et al. 2021 NHESS", + "content": "Kruspe, A., Kersten, J., & Klan, F. (2021). Review article: Detection of actionable tweets in crisis events. Natural Hazards Earth System Science., 21, 1825–1845. https://doi.org/10.5194/nhess-21-1825-2021.", + "attributes": [ + { + "attribute": "source", + "value": "NHESS (2021) 21:1825-1845" + }, + { + "attribute": "authors", + "value": "Kruspe A.; Kersten J.; Klan F." + }, + { + "attribute": "year", + "value": "2021" + }, + { + "attribute": "title", + "value": "Detection of actionable tweets in crisis events" + }, + { + "attribute": "venue", + "value": "Natural Hazards Earth System Science" + }, + { + "attribute": "pages", + "value": "1825-1845" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.5194/nhess-21-1825-2021" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Convolutions depth: Szegedy 2015 Going Deeper with Convolutions (repeat)", + "content": "Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., & Rabinovich, A. (2015). Going deeper with convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 1-9).", + "attributes": [ + { + "attribute": "source", + "value": "CVPR 2015" + }, + { + "attribute": "authors", + "value": "Szegedy C.; Liu W.; Jia Y.; Sermanet P.; Reed S.; Anguelov D.; Rabinovich A." + }, + { + "attribute": "year", + "value": "2015" + }, + { + "attribute": "title", + "value": "Going deeper with convolutions" + }, + { + "attribute": "venue", + "value": "CVPR 2015" + }, + { + "attribute": "pages", + "value": "1-9" + }, + { + "attribute": "doi", + "value": "N/A" + }, + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "confidence", + "value": "low" + } + ] + }, + { + "type": "comment", + "insight": "Note: The References page shows formatting inconsistencies and duplicated entries (e.g., Chaaraoui 2014 repeated, typos in author names).", + "content": "Chaaraoui, A. A., Padilla-López, J. R., Ferrández-Pastor, F. J., Nieto-Hidalgo, M., & Flórez-Revuelta, F. (2014). A vision-based system for intelligent monitoring: Human behaviour analysis and privacy by context. A Vision-Based System for Intelligent Monitoring: Human Behaviour Analysis and Privacy by Context. Alexandro, 14(5), 8895–8925. https://doi.org/10.3390/s140508895", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "note", + "value": "Formatting irregularities and duplication observed on page." + } + ] + }, + { + "type": "fact", + "insight": "Additional classic computer vision references appearing in the list: R-CNN family, FPN, YOLOv2, etc.", + "content": "The references include foundational works in deep learning and object detection, such as Fast R-CNN (Girshick 2015), Feature Pyramid Networks (Lin et al. 2017), Faster R-CNN (Ren et al. 2015), and Going Deeper with Convolutions (Szegedy et al. 2015).", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "note", + "value": "Synthesis from multiple cited works: Fast R-CNN, FPN, Faster R-CNN, etc." + } + ] + }, + { + "type": "fact", + "insight": "Chaaraoui et al. (2014) introduce a vision-based system for intelligent monitoring focusing on human behaviour analysis and context-based privacy.", + "content": "Chaaraoui, A. A., Padilla-López, J. R., Ferrández-Pastor, F. J., Nieto-Hidalgo, M., & Flórez-Revuelta, F. (2014). A vision-based system for intelligent monitoring: Human behaviour analysis and privacy by context. A Vision-Based System for Intelligent Monitoring: Human Behaviour Analysis and Privacy by Context Alexandros,14(5), 8895–8925. https://doi.org/10.3390/s140508895", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "year", + "value": "2014" + }, + { + "attribute": "authors", + "value": "Chaaraoui A A; Padilla-López J R; Ferrández-Pastor F J; Nieto-Hidalgo M; Flórez-Revuelta F" + }, + { + "attribute": "title", + "value": "A vision-based system for intelligent monitoring: Human behaviour analysis and privacy by context" + }, + { + "attribute": "venue", + "value": "Sensors 2014; 14(5); 8895-8925" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.3390/s140508895" + } + ] + }, + { + "type": "fact", + "insight": "Mishra & Gupta (2017) provide an overview of deep machine learning and neural networks in the IAES IJ-AI journal.", + "content": "Mishra, C., & Gupta, D. L. (2017). Deep Machine Learning and Neural Networks: An Overview. IAES International Journal of Artificial Intelligence (IJ-AI), 6(2), 66.", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "year", + "value": "2017" + }, + { + "attribute": "authors", + "value": "Mishra C; Gupta DL" + }, + { + "attribute": "title", + "value": "Deep Machine Learning and Neural Networks: An Overview" + }, + { + "attribute": "venue", + "value": "IAES IJ-AI 6(2) 2017" + }, + { + "attribute": "doi", + "value": "N/A" + } + ] + }, + { + "type": "fact", + "insight": "Tee et al. (2015) discuss an activity monitoring system for the elderly in the ARPN Journal of Engineering and Applied Sciences.", + "content": "Tee, K. S., Zulkifli, A. H. B., & Soon, C. F. (2015). An activity monitoring system for elderly. ARPN Journal of Engineering and Applied Sciences, 10 (18), 8467–8472.", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "year", + "value": "2015" + }, + { + "attribute": "authors", + "value": "Tee K S; Zulkifli A H B; Soon C F" + }, + { + "attribute": "title", + "value": "An activity monitoring system for elderly" + }, + { + "attribute": "venue", + "value": "ARPN J Eng Appl Sci 10(18) 2015" + }, + { + "attribute": "doi", + "value": "N/A" + } + ] + }, + { + "type": "fact", + "insight": "Wang (2019) defines artificial intelligence in a Journal of Artificial General Intelligence article.", + "content": "Wang, P. (2019). On Defining Artificial Intelligence. Journal of Artificial General Intelligence , 10 (2), 1–37. https://doi.org/10.2478/jagi-2019-0002", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "year", + "value": "2019" + }, + { + "attribute": "authors", + "value": "Wang P." + }, + { + "attribute": "title", + "value": "On Defining Artificial Intelligence" + }, + { + "attribute": "venue", + "value": "Journal of Artificial General Intelligence 10(2) 2019" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.2478/jagi-2019-0002" + } + ] + }, + { + "type": "fact", + "insight": "Attaran & Deb (2018) discuss machine learning as a new driver for competitive advantage in knowledge engineering and data mining.", + "content": "Attaran, M., & Deb, P. (2018). Machine Learning : The New ’ Big Thing ’ for Competitive Advantage. International Journal of Knowledge Engineering and Data Mining, 5(4), 277-305. https://doi.org/10.1504/IJKEDM.2018.10015621.", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "year", + "value": "2018" + }, + { + "attribute": "authors", + "value": "Attaran M; Deb P" + }, + { + "attribute": "title", + "value": "Machine Learning: The New ’ Big Thing ’ for Competitive Advantage" + }, + { + "attribute": "venue", + "value": "Int J Knowl Eng Data Mining 5(4) 2018" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.1504/IJKEDM.2018.10015621" + } + ] + }, + { + "type": "fact", + "insight": "Deng et al. (2009) describe the ImageNet database, a large-scale hierarchical image database, presented at CVPR 2009.", + "content": "Deng, J., Dong, W., Socher, R., Li, L. J., Li, K., & Fei-Fei, L. (2009). Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition (pp. 248-255).", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "year", + "value": "2009" + }, + { + "attribute": "authors", + "value": "Deng J; Dong W; Socher R; Li LJ; Li K; Fei-Fei L" + }, + { + "attribute": "title", + "value": "Imagenet: A large-scale hierarchical image database" + }, + { + "attribute": "venue", + "value": "CVPR 2009" + } + ] + }, + { + "type": "fact", + "insight": "Forbes et al. (2020) explore WiFi-based human activity recognition using Raspberry Pi at ICTAI 2020.", + "content": "Forbes, G., Massie, S., & Craw, S. (2020) WiFi-based Human Activity Recognition using Raspberry Pi. International Conference on Tools with Artificial Intelligence (ICTAI), 5(3), 722–730. https://doi.org/10.1016/j.ijtst.2016.12.001", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "year", + "value": "2020" + }, + { + "attribute": "authors", + "value": "Forbes G; Massie S; Craw S" + }, + { + "attribute": "title", + "value": "WiFi-based Human Activity Recognition using Raspberry Pi" + }, + { + "attribute": "venue", + "value": "ICTAI 2020 5(3) 722-730" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.1016/j.ijtst.2016.12.001" + } + ] + }, + { + "type": "fact", + "insight": "Girshick (2015) introduced Fast R-CNN for object detection at CVPR/ICCV venues.", + "content": "Girshick, R. (2015). Fast r-cnn. In Proceedings of the IEEE international conference on computer vision (pp. 1440-1448).", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "year", + "value": "2015" + }, + { + "attribute": "authors", + "value": "Girshick R." + }, + { + "attribute": "title", + "value": "Fast R-CNN" + }, + { + "attribute": "venue", + "value": "IEEE CVPR/ICCV 2015" + } + ] + }, + { + "type": "fact", + "insight": "Gupta & Mishra (2017) provide another overview of deep machine learning and neural networks in the IAES IJ-AI journal.", + "content": "Gupta, D. L., & Mishra, C. (2017). Deep Machine Learning and Neural Networks: An Overview. IAES International Journal of Artificial Intelligence (IJ-AI), 6(2), 66-72.", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "year", + "value": "2017" + }, + { + "attribute": "authors", + "value": "Gupta D L; Mishra C" + }, + { + "attribute": "title", + "value": "Deep Machine Learning and Neural Networks: An Overview" + }, + { + "attribute": "venue", + "value": "IAES IJ-AI 6(2) 2017" + }, + { + "attribute": "doi", + "value": "N/A" + } + ] + }, + { + "type": "fact", + "insight": "Chaaraoui et al. (2014) appear again with a diacritics-formatted version of the same title.", + "content": "Chaaraoui, A. A., Padilla-L´opez, J. R., Ferr´andez-Pastor, F. J., Nieto-Hidalgo, M., & Fl´orez-Revuelta, F. (2014). A Vision-Based System for Intelligent Monitoring: Human Behaviour Analysis and Privacy by Context. Sensors 2014, 14, 8895-8925.", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "year", + "value": "2014" + }, + { + "attribute": "authors", + "value": "Chaaraoui A A; Padilla-López J R; Ferrández-Pastor F J; Nieto-Hidalgo M; Flórez-Revuelta F" + }, + { + "attribute": "title", + "value": "A Vision-Based System for Intelligent Monitoring: Human Behaviour Analysis and Privacy by Context" + }, + { + "attribute": "venue", + "value": "Sensors 2014; 14; 8895-8925" + }, + { + "attribute": "doi", + "value": "N/A" + } + ] + }, + { + "type": "fact", + "insight": "Deng et al. (2009)Über containment of ImageNet entry; duplicate line included for completeness.", + "content": "Deng, J., Dong, W., Socher, R., Li, L. J., Li, K., & Fei-Fei, L. (2009). Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition (pp. 248-255).", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "year", + "value": "2009" + }, + { + "attribute": "authors", + "value": "Deng J; Dong W; Socher R; Li LJ; Li K; Fei-Fei L" + }, + { + "attribute": "title", + "value": "Imagenet: A large-scale hierarchical image database" + }, + { + "attribute": "venue", + "value": "CVPR 2009" + } + ] + }, + { + "type": "fact", + "insight": "Forbes et al. (2020) entry connects with a 2020 ICTAI publication on WiFi-based activity recognition.", + "content": "Forbes, G., Massie, S., & Craw, S. (2020) WiFi-based Human Activity Recognition using Raspberry Pi. International Conference on Tools with Artificial Intelligence (ICTAI), 5(3), 722–730. https://doi.org/10.1016/j.ijtst.2016.12.001", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "year", + "value": "2020" + }, + { + "attribute": "authors", + "value": "Forbes G; Massie S; Craw S" + }, + { + "attribute": "title", + "value": "WiFi-based Human Activity Recognition using Raspberry Pi" + }, + { + "attribute": "venue", + "value": "ICTAI 2020 5(3) 722-730" + }, + { + "attribute": "doi", + "value": "https://doi.org/10.1016/j.ijtst.2016.12.001" + } + ] + }, + { + "type": "fact", + "insight": "Girshick (2015) describes Fast R-CNN as an object detection method.", + "content": "Girshick, R. (2015). Fast r-cnn. In Proceedings of the IEEE international conference on computer vision (pp. 1440-1448).", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "year", + "value": "2015" + }, + { + "attribute": "authors", + "value": "Girshick R." + }, + { + "attribute": "title", + "value": "Fast R-CNN" + }, + { + "attribute": "venue", + "value": "CVPR/ICCV 2015" + } + ] + }, + { + "type": "fact", + "insight": "Gupta & Mishra (2017) present another overview of deep machine learning and neural networks in the IAES IJ-AI journal.", + "content": "Gupta, D. L., & Mishra, C. (2017). Deep Machine Learning and Neural Networks: An Overview. IAES International Journal of Artificial Intelligence (IJ-AI), 6(2), 66-72.", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "year", + "value": "2017" + }, + { + "attribute": "authors", + "value": "Gupta D L; Mishra C" + }, + { + "attribute": "title", + "value": "Deep Machine Learning and Neural Networks: An Overview" + }, + { + "attribute": "venue", + "value": "IAES IJ-AI 6(2) 2017" + }, + { + "attribute": "doi", + "value": "N/A" + } + ] + }, + { + "type": "comment", + "insight": "The page appears to be a References section focused on deep learning methods for traffic monitoring, vehicle detection, and tracking.", + "content": "Traffic Monitoring System. sustainability, 12, 9177, 1-21 .", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "page", + "value": "55" + }, + { + "attribute": "topic", + "value": "Deep Learning for Traffic Monitoring" + } + ] + }, + { + "type": "fact", + "insight": "Faster R-CNN is cited as a real-time object detection method (Ren et al., 2015).", + "content": "Ren, S., He, K., Girshick, R., & Sun, J. (2015). Faster r-cnn: Towards real-time object detection with region proposal networks. Advancements in neural information processing systems, 28.", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "topic", + "value": "Real-time object detection" + }, + { + "attribute": "year", + "value": "2015" + } + ] + }, + { + "type": "fact", + "insight": "The classic convolutional network paper Going deeper with convolutions (Szegedy et al., 2015) is referenced, indicating reliance on deeper CNN architectures.", + "content": "Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., & Rabinovich, A. (2015). Going deeper with convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 1-9).", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "topic", + "value": "CNN architectures" + } + ] + }, + { + "type": "fact", + "insight": "The list includes P-VANET: Deep but lightweight networks for real-time object detection (Kim et al., 2016), pointing to lightweight real-time detectors.", + "content": "Kim, K. H., Hong, S., Roh, B., Cheon, Y., & Park, M. (2016). Pvanet: Deep but lightweight neural networks for real-time object detection. arXiv preprint arXiv:1608.08021.", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "topic", + "value": "Lightweight real-time detectors" + } + ] + }, + { + "type": "fact", + "insight": "YOLOv2-based vehicle tracking with GIS-based algorithms is cited (Malaainine et al., 2021), illustrating the use of fast real-time detection in geospatial contexts.", + "content": "Malaainine, M. E, I., Lechgar, H., & Rhinane, H. (2021). Y OLOv2 Deep Learning Model and GIS Based Algorithms for Vehicle Tracking. Journal of Geographic Information System, 2021, 13, 395-409 https://doi.org/10.4236/jgis.2021.134022.", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "topic", + "value": "YOLOv2 for vehicle tracking" + } + ] + }, + { + "type": "fact", + "insight": "A set of reviews on deep learning and AI-enabled systems, including human-centered ML and big data analytics, is cited (e.g., Najafabadi et al., 2015; Sarker, 2021).", + "content": "Najafabadi, M. M., Villanustre, F., Khoshgoftaar, T. M., Seliya, N., Wald, R., & Muharemagic, E. (2015). Deep learning applications and challenges in big data analytics. Journal of Big Data, 1–21. https://doi.org/10.1186/s40537-014-0007-7; Sarker, I. H. (2021). Machine Learning : Algorithms , Real ‑ World Applications and Research Directions. SN Computer Science, 2 (3), 1–21. https://doi.org/10.1007/s42979-021-00592-x", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "topic", + "value": "Deep learning in big data and general ML" + } + ] + }, + { + "type": "fact", + "insight": "Deep learning-based vehicle tracking in traffic management is represented (Shehata et al., 2020).", + "content": "Shehata, M., Abo-alez, R., Zaghlool, F., & Abou-kreisha, M. T. (2020). Deep Learning Based Vehicle Tracking in Traffic Management. International Journal of Computer Trends and Technology (IJCTT), 67(3), 5-8. https://doi.org/10.14445/22312803/IJCTT-V67I3P102", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "topic", + "value": "Vehicle tracking in traffic management" + } + ] + }, + { + "type": "fact", + "insight": "There are references to smart monitoring with Raspberry Pi and smartphones (Surya & Ningsih, 2019), indicating use of affordable IoT hardware for surveillance.", + "content": "Surya, E., & Ningsih, Y.K. (2019). Smart Monitoring System Using Raspberry-Pi and Smartphone. Department of Electrical Engineering Faculty of Industrial Technology , Trisakti University Jakarta Indonesia, 7(1), 72-84.", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "topic", + "value": "Smart monitoring with Raspberry Pi" + } + ] + }, + { + "type": "fact", + "insight": "The page includes the classic 'visual detection' and 'deep learning' references such as Zhang et al. (2020) and Zhou et al. (2020) on vehicle detection/tracking.", + "content": "Zhang, Y., Song, X.,Wang, M., Guan, T., Liu, J., Wang, Z., Zhen, Y., Zhang, D., & Gu, Xiaoyi. (2020). Research on visual vehicle detection and tracking based on deep learning. IOP Conference Series: Materials Science and Engineering, 892(2020), 1-7 https://doi.org/10.1088/1742-6596/1621/1/012048.", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "topic", + "value": "Visual vehicle detection and tracking" + } + ] + }, + { + "type": "fact", + "insight": "A 2020 study on deep learning vehicle tracking is cited (Zhou et al., 2020).", + "content": "Zhou, Y., Zhou, J., & Liao, F. (2020). Research on Vehicle Tracking Algorithm Based on Deep Learning. Journal of Physics: Conference Series, 1621 (2020), 1-8. https://doi.org/10.1088/1742-6596/1621/1/012048.", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "topic", + "value": "Vehicle tracking algorithm" + } + ] + }, + { + "type": "fact", + "insight": "This page is a References/Bibliography section listing works related to traffic monitoring systems and deep learning-based vehicle detection/tracking.", + "content": "Traffic Monitoring System. sustainability, 12, 9177, 1-21 .", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "topic", + "value": "Traffic Monitoring System" + }, + { + "attribute": "page", + "value": "56" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Ren, S., He, K., Girshick, R., & Sun, J. (2015). Faster R-CNN: Towards real-time object detection with region proposal networks.", + "content": "Ren, S., He, K., Girshick, R., & Sun, J. (2015). Faster r-cnn: Towards real-time object detection with region proposal networks. Advances in neural information processing systems, 28.", + "attributes": [ + { + "attribute": "author", + "value": "Ren et al." + }, + { + "attribute": "date", + "value": "2015" + }, + { + "attribute": "title", + "value": "Faster R-CNN: Towards real-time object detection with region proposal networks" + }, + { + "attribute": "venue", + "value": "Advances in Neural Information Processing Systems (NIPS/NeurIPS)" + } + ] + }, + { + "type": "fact", + "insight": "Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., & Rabinovich, A. (2015). Going deeper with convolutions.", + "content": "Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., & Rabinovich, A. (2015). Going deeper with convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 1-9).", + "attributes": [ + { + "attribute": "author", + "value": "Szegedy et al." + }, + { + "attribute": "date", + "value": "2015" + }, + { + "attribute": "title", + "value": "Going deeper with convolutions" + }, + { + "attribute": "venue", + "value": "IEEE Conference on Computer Vision and Pattern Recognition (CVPR)" + }, + { + "attribute": "pages", + "value": "pp. 1-9" + } + ] + }, + { + "type": "fact", + "insight": "Najafabadi, M. M., Villanustre, F., Khoshgoftaar, T. M., Seliya, N., Wald, R., & Muharemagic, E. (2015). Deep learning applications and challenges in big data analytics.", + "content": "Najafabadi, M. M., Villanustre, F., Khoshgoftaar, T. M., Seliya, N., Wald, R., & Muharemagic, E. (2015). Deep learning applications and challenges in big data analytics. Journal of Big Data, 1–21.", + "attributes": [ + { + "attribute": "author", + "value": "Najafabadi et al." + }, + { + "attribute": "date", + "value": "2015" + }, + { + "attribute": "title", + "value": "Deep learning applications and challenges in big data analytics" + }, + { + "attribute": "venue", + "value": "Journal of Big Data" + } + ] + }, + { + "type": "fact", + "insight": "Surya, E., & Ningsih, Y.K. (2019). Smart Monitoring System Using Raspberry-Pi and Smartphone.", + "content": "Surya, E., & , Ningsih, Y.K. (2019). Smart Monitoring System Using Raspberry-Pi and Smartphone.", + "attributes": [ + { + "attribute": "author", + "value": "Surya & Ningsih (2019)" + }, + { + "attribute": "date", + "value": "2019" + }, + { + "attribute": "title", + "value": "Smart Monitoring System Using Raspberry-Pi and Smartphone" + } + ] + }, + { + "type": "fact", + "insight": "Shehata, M., Abo-alez, R., Zaghlool, F., & Abou-kreisha, M. T. (2020). Deep Learning Based Vehicle Tracking in Traffic Management.", + "content": "Shehata, M., Abo-alez, R., Zaghlool, F., & Abou-kreisha, M. T. (2020). Deep Learning Based Vehicle Tracking in Traffic Management. International Journal of Computer Trends and Technology (IJCTT), 67(3), 5-8.", + "attributes": [ + { + "attribute": "author", + "value": "Shehata et al." + }, + { + "attribute": "date", + "value": "2020" + }, + { + "attribute": "title", + "value": "Deep Learning Based Vehicle Tracking in Traffic Management" + }, + { + "attribute": "venue", + "value": "IJCTT" + } + ] + }, + { + "type": "fact", + "insight": "Usmankhujaev, S., Baydadaev, S., & Woo, K. J. (2020). Real-Time, Deep Learning Based Wrong Direction Detection.", + "content": "Usmankhujaev, S., Baydadaev, S., & Woo, K. J. (2020). Real-Time, Deep Learning Based Wrong Direction Detection. Applied Sciences, 10(2020), 1-13.", + "attributes": [ + { + "attribute": "author", + "value": "Usmankhujaev et al." + }, + { + "attribute": "date", + "value": "2020" + }, + { + "attribute": "title", + "value": "Real-Time, Deep Learning Based Wrong Direction Detection" + }, + { + "attribute": "venue", + "value": "Applied Sciences" + } + ] + }, + { + "type": "fact", + "insight": "Wang, P. (2019). On Defining Artificial Intelligence.", + "content": "Wang, P. (2019). On Defining Artificial Intelligence. Journal of Artificial General Intelligence 10(2) 1-37, 2019.", + "attributes": [ + { + "attribute": "author", + "value": "Wang" + }, + { + "attribute": "date", + "value": "2019" + }, + { + "attribute": "title", + "value": "On Defining Artificial Intelligence" + }, + { + "attribute": "venue", + "value": "Journal of Artificial General Intelligence" + } + ] + }, + { + "type": "fact", + "insight": "Zhang, Y., Song, X., Wang, M., Guan, T., Liu, J., Wang, Z., Zhen, Y., Zhang, D., & Gu, Xiaoyi. (2020). Research on visual vehicle detection and tracking based on deep learning.", + "content": "Zhang, Y., Song, X.,Wang, M., Guan, T., Liu, J., Wang, Z., Zhen, Y., Zhang, D., & Gu, Xiaoyi. (2020). Research on visual vehicle detection and tracking based on deep learning. IOP Conference Series: Materials Science and Engineering, 892(2020), 1-7.", + "attributes": [ + { + "attribute": "author", + "value": "Zhang et al." + }, + { + "attribute": "date", + "value": "2020" + }, + { + "attribute": "title", + "value": "Research on visual vehicle detection and tracking based on deep learning" + }, + { + "attribute": "venue", + "value": "IOP Conference Series: Materials Science and Engineering" + } + ] + }, + { + "type": "fact", + "insight": "Zhou, Y., Zhou, J., & Liao, F. (2020). Research on Vehicle Tracking Algorithm Based on Deep Learning.", + "content": "Zhou, Y., Zhou, J., & Liao, F. (2020). Research on Vehicle Tracking Algorithm Based on Deep Learning. Journal of Physics: Conference Series, 1621 (2020), 1-8.", + "attributes": [ + { + "attribute": "author", + "value": "Zhou et al." + }, + { + "attribute": "date", + "value": "2020" + }, + { + "attribute": "title", + "value": "Research on Vehicle Tracking Algorithm Based on Deep Learning" + }, + { + "attribute": "venue", + "value": "Journal of Physics: Conference Series" + } + ] + }, + { + "type": "fact", + "insight": "https://doi.org/10.1186/s40537-014-0007-7 is a DOI listed in the references.", + "content": "https://doi.org/10.1186/s40537-014-0007-7", + "attributes": [ + { + "attribute": "type", + "value": "DOI" + }, + { + "attribute": "value", + "value": "https://doi.org/10.1186/s40537-014-0007-7" + } + ] + }, + { + "type": "comment", + "insight": "Formatting irregularities are present in the reference list (e.g., Surya entry shows an extra comma; spaces and punctuation are inconsistently applied).", + "content": "Surya, E., & , Ningsih, Y.K. (2019). Smart Monitoring System Using Raspberry-Pi and Smartphone.", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "note", + "value": "Formatting irregularities" + } + ] + }, + { + "type": "fact", + "insight": "The page lists a mix of journals, conferences, and conference series (e.g., IJCTT, Applied Sciences, IOP Conference Series, SN Computer Science); indicating a broad literature base.", + "content": "Shehata, M., Abo-alez, R., Zaghlool, F., & Abou-kreisha, M. T. (2020). Deep Learning Based Vehicle Tracking in Traffic Management. IJCTT, 67(3), 5-8; Surya entry; Zhang et al. 2020 IOP Conference Series; Zhou et al. 2020 JPCS.", + "attributes": [ + { + "attribute": "section", + "value": "References" + }, + { + "attribute": "note", + "value": "Diversity of venues (IJCTT, Applied Sciences, IOP, JPCS)" + } + ] + }, + { + "type": "fact", + "insight": "This is the final page of the document (page 56).", + "content": "56", + "attributes": [ + { + "attribute": "section", + "value": "Page footer" + }, + { + "attribute": "page", + "value": "56" + } + ] + } + ] + }, + "EXTROPIC PAPER": { + "author_metadata": "Extropic, a company focused on developing innovative technologies for sustainable computing.", + "source_metadata": "Extropic Corporation research paper", + "knowledge": [ + { + "type": "fact", + "insight": "The paper proposes an all-transistor probabilistic computer architecture called Denoising Thermodynamic Models (DTMs) that could achieve performance parity with GPUs while using approximately 10,000 times less energy on simple image benchmarks.", + "content": "Asystem-levelanalysisindicatesthatdevicesbased on our architecture could achieve performance parity with GPUs on a simple image benchmark using approximately 10,000 times less energy.", + "attributes": [ + { + "attribute": "source", + "value": "Extropic Corporation research paper" + }, + { + "attribute": "date", + "value": "October 29, 2025" + }, + { + "attribute": "energy_efficiency_claim", + "value": "10,000x improvement" + }, + { + "attribute": "benchmark_type", + "value": "simple image benchmark" + } + ] + }, + { + "type": "fact", + "insight": "U.S. AI data centers could consume 10% of all U.S. energy production by 2030, and current annual spending exceeds the inflation-adjusted cost of the Apollo program.", + "content": "Every year, U.S. firms spend an amount larger than the inflation-adjusted cost of the Apollo program on AI-focused data centers [1, 2]. By 2030, these data centers could consume 10% of all of the energy produced in the U.S. [3].", + "attributes": [ + { + "attribute": "source", + "value": "U.S. energy consumption projections" + }, + { + "attribute": "date", + "value": "2025 projection" + }, + { + "attribute": "energy_consumption", + "value": "10% of U.S. energy by 2030" + }, + { + "attribute": "annual_spending", + "value": "Exceeds Apollo program cost" + } + ] + }, + { + "type": "opinion", + "insight": "Current AI algorithms have evolved in a suboptimal direction due to the 'Hardware Lottery' problem, where hardware availability has influenced algorithm research.", + "content": "Had a different style of hardware been popular in the last few decades, AI algorithms would have evolved in a completely different direction, and possibly a more energy-efficient one. This interplay between algorithm research and hardware availability is known as the 'Hardware Lottery' [19], and it entrenches hardware-algorithm pairings that may be far from optimal.", + "attributes": [ + { + "attribute": "source", + "value": "Research analysis" + }, + { + "attribute": "problem", + "value": "Hardware Lottery" + }, + { + "attribute": "consequence", + "value": "Suboptimal algorithm evolution" + }, + { + "attribute": "sentiment", + "value": "Critical" + } + ] + }, + { + "type": "fact", + "insight": "Existing probabilistic computer approaches for EBMs suffer from the 'mixing-expressivity tradeoff' (MET), where better modeling performance leads to exponentially longer mixing times and higher energy costs.", + "content": "The mixing-expressivity tradeoff (MET) summarizes this issue with existing probabilistic computer architectures, reflecting the fact that modeling performance and sampling hardness are coupled for MEBMs. Specifically, as the expressivity (modeling performance) of an MEBM increases, its mixing time (the amount of computational effort needed to draw independent samples from the MEBM's distribution) becomes progressively longer, resulting in expensive inference and unstable training [52, 53].", + "attributes": [ + { + "attribute": "problem", + "value": "Mixing-Expressivity Tradeoff (MET)" + }, + { + "attribute": "consequence", + "value": "Exponentially longer mixing times" + }, + { + "attribute": "impact", + "value": "Expensive inference and unstable training" + }, + { + "attribute": "solution_type", + "value": "Addressed by DTM approach" + } + ] + }, + { + "type": "fact", + "insight": "The DTCA uses all-transistor hardware with subthreshold transistor dynamics for random number generation, avoiding exotic components and enabling CMOS scalability.", + "content": "To enable a near-term, large-scale realization of the DTCA, we leveraged the shot-noise dynamics of subthreshold transistors [45] to build an RNG that is fast, energy-efficient, and small. Our all-transistor RNG is programmable and has the desired sigmoidal response to a control voltage, as shown by experimental measurements in Fig. 4 (a).", + "attributes": [ + { + "attribute": "hardware_approach", + "value": "All-transistor implementation" + }, + { + "attribute": "technology", + "value": "Subthreshold transistor dynamics" + }, + { + "attribute": "scalability", + "value": "CMOS compatible" + }, + { + "attribute": "component_type", + "value": "RNG (Random Number Generator)" + } + ] + }, + { + "type": "opinion", + "insight": "DTMs represent the first scalable method for applying probabilistic hardware to machine learning by chaining multiple EBMs to gradually build complexity.", + "content": "At the top level, we introduce a new probabilistic computer architecture that runs Denoising Thermodynamic Models (DTMs) instead of monolithic EBMs. As their name suggests, rather than using the hardware's EBM to model data distributions directly, DTMs sequentially compose many hardware EBMs to model a process that denoises the data gradually. Diffusion models [18, 44] also follow this denoising procedure and are much more capable than EBMs. This key architectural change addresses a fundamental issue with previous approaches and represents the first scalable method for applying probabilistic hardware to machine learning.", + "attributes": [ + { + "attribute": "innovation", + "value": "First scalable probabilistic hardware approach" + }, + { + "attribute": "key_change", + "value": "Sequential EBM composition" + }, + { + "attribute": "benefit", + "value": "Avoids mixing-expressivity tradeoff" + }, + { + "attribute": "significance", + "value": "Fundamental architectural breakthrough" + } + ] + }, + { + "type": "comment", + "insight": "The research specifically uses sparse Boltzmann machines (Ising models) as the EBM implementation due to their hardware efficiency and simple Gibbs sampling update rules.", + "content": "The DTM that produced the results shown in Fig. 1 used Boltzmann machine EBMs. Boltzmann machines, also known as Ising models in physics, use binary random variables and are the simplest type of discrete-variable EBM.\n\nBoltzmann machines are hardware efficient because the Gibbs sampling update rule required to sample from them is simple. Boltzmann machines implement energy functions of the form\nE(x) =−β \n⟨\n∑_{i≠j} x_i J_ij x_j + ∑_{i=1}^n h_i x_i\n⟩\n,(10)", + "attributes": [ + { + "attribute": "implementation", + "value": "Sparse Boltzmann machines" + }, + { + "attribute": "model_type", + "value": "Ising models/EBMs" + }, + { + "attribute": "efficiency_reason", + "value": "Simple Gibbs sampling" + }, + { + "attribute": "variable_type", + "value": "Binary random variables" + } + ] + }, + { + "type": "fact", + "insight": "GPU performance per joule doubles every few years, making it extremely difficult for new computing schemes to achieve mainstream adoption despite theoretical advantages.", + "content": "In addition to these integration challenges, GPU performance per joule is doubling every few years [24], making it very difficult for cutting-edge computing schemes to gain mainstream adoption.", + "attributes": [ + { + "attribute": "challenge", + "value": "GPU efficiency improvement rate" + }, + { + "attribute": "rate", + "value": "Doubling every few years" + }, + { + "attribute": "impact", + "value": "Barriers to new adoption" + }, + { + "attribute": "context", + "value": "Competitive landscape analysis" + } + ] + }, + { + "type": "comment", + "insight": "The DTCA architecture can be implemented in various modular configurations, including distinct physical circuitry on the same chip, across multiple communicating chips, or reprogrammed hardware with different weights.", + "content": "The modular nature of DTMs enables various hardware implementations. For example, each EBM in the chain can be implemented using distinct physical circuitry on the same chip, as shown in Fig. 3 (b). Alternatively, the various EBMs may be split across several communicating chips or implemented by the same hardware, reprogrammed with distinct sets of weights at different times.", + "attributes": [ + { + "attribute": "architecture", + "value": "Modular DTCA" + }, + { + "attribute": "implementation_options", + "value": "Multiple chip configurations" + }, + { + "attribute": "flexibility", + "value": "Reprogrammable hardware" + }, + { + "attribute": "scalability", + "value": "Various deployment options" + } + ] + }, + { + "type": "fact", + "insight": "The mixing-expressivity tradeoff (MET) summarizes the coupling between modeling performance and sampling hardness for MEBMs, where increased expressivity leads to longer mixing times and more expensive inference.", + "content": "The mixing-expressivity tradeoff (MET) summarizes this issue with existing probabilistic computer architectures, reflecting the fact that modeling performance and sampling hardness are coupled for MEBMs. Specifically, as the expressivity (modeling performance) of an MEBM increases, its mixing time (the amount of computational effort needed to draw independent samples from the MEBM's distribution) becomes progressively longer, resulting in expensive inference and unstable training [52, 53].", + "attributes": [ + { + "attribute": "source", + "value": "Academic paper text" + }, + { + "attribute": "reference", + "value": "[52, 53]" + } + ] + }, + { + "type": "fact", + "insight": "DTMs (Denoising Thermodynamic Models) merge EBMs with diffusion models to provide an alternative probabilistic computing approach that addresses the MET.", + "content": "DTMs merge EBMs with diffusion models, offering an alternative path for probabilistic computing that assuages the MET. DTMs are a slight generalization of recent work from deep learning practitioners that has pushed the frontier of EBM performance [57–60].", + "attributes": [ + { + "attribute": "source", + "value": "Academic paper text" + }, + { + "attribute": "reference", + "value": "[57–60]" + } + ] + }, + { + "type": "fact", + "insight": "The forward process in denoising diffusion models follows a Markov chain structure and has a unique stationary distribution with a simple form.", + "content": "Denoising models attempt to reverse a process that gradually transforms the data distribution Q(x0) into simple noise. This forward process is given by the Markov chain Q(x0, . . . , xT ) = Q(x0) ∏t=1T Q(xt|xt−1). (3) The forward process is typically chosen such that it has a unique stationary distribution Q(xT ), which takes a simple form (e.g., Gaussian or uniform).", + "attributes": [ + { + "attribute": "source", + "value": "Academic paper text" + }, + { + "attribute": "equation", + "value": "Eq. (3)" + } + ] + }, + { + "type": "opinion", + "insight": "MEBMs have a fundamental flaw that makes them energetically costly to scale at the probabilistic computing level.", + "content": "The MET makes it clear that MEBMs have a flaw that makes them challenging and energetically costly to scale.", + "attributes": [ + { + "attribute": "source", + "value": "Author's assessment" + }, + { + "attribute": "perspective", + "value": "Critical" + } + ] + }, + { + "type": "opinion", + "insight": "DTMs successfully overcome the mixing-expressivity tradeoff by using a gradual complexity building approach.", + "content": "Instead of trying to use a single EBM to model the data, DTMs chain many EBMs to gradually build up to the complexity of the data distribution. This gradual buildup of complexity allows the landscape of each EBM in the chain to remain relatively simple (and easy to sample) without limiting the complexity of the distribution modeled by the chain as a whole;", + "attributes": [ + { + "attribute": "source", + "value": "Author's technical assessment" + }, + { + "attribute": "perspective", + "value": "Positive/optimistic" + } + ] + }, + { + "type": "comment", + "insight": "The mixing-expressivity tradeoff creates energy barriers between data modes that make sampling computationally expensive.", + "content": "For large differences in energy, like those encountered when trying to move between two valleys separated by a significant barrier, this probability can be very close to zero. These barriers grind the iterative sampler to a halt.", + "attributes": [ + { + "attribute": "source", + "value": "Technical description" + }, + { + "attribute": "metaphor", + "value": "Valley/barrier analogy" + } + ] + }, + { + "type": "comment", + "insight": "Denoising diffusion models work by reversing a gradual noise addition process to generate data from simple noise.", + "content": "Reversal of the forward process is achieved by learning a set of distributions Pθ(xt−1|xt) that approximate the reversal of each conditional in Eq. (3). In doing so, we learn a map from simple noise to the data distribution, which can then be used to generate new data.", + "attributes": [ + { + "attribute": "source", + "value": "Methodological description" + }, + { + "attribute": "process", + "value": "Reverse diffusion" + } + ] + }, + { + "type": "comment", + "insight": "Traditional diffusion models use sufficiently fine-grained forward processes for their implementation.", + "content": "In traditional diffusion models, the forward process is made to be sufficiently fine-grained (using a large num", + "attributes": [ + { + "attribute": "source", + "value": "Comparative statement about traditional approaches" + }, + { + "attribute": "completeness", + "value": "Incomplete sentence" + } + ] + }, + { + "type": "fact", + "insight": "Traditional diffusion models use a neural network parameterized conditional distribution (Gaussian or categorical) to approximate the reverse process, minimizing KL divergence between joint distributions Q and Pθ", + "content": "ber of stepsT) such that the conditional distribution of each step in the reverse process takes some simple form (such as Gaussian or categorical). This simple distribution is parameterized by a neural network, which is then trained to minimize the Kullback-Leibler (KL) divergence between the joint distributionsQandP θ, LDN (θ) =D Q(x0, . . . , xT ) Pθ(x0, . . . , xT ) ,(4)", + "attributes": [ + { + "attribute": "method", + "value": "traditional diffusion models" + }, + { + "attribute": "objective", + "value": "minimize KL divergence" + }, + { + "attribute": "section", + "value": "model training" + } + ] + }, + { + "type": "fact", + "insight": "The joint distribution of diffusion models is the product of learned conditionals: Pθ(x0, ..., xT) = Q(xT) ∏_{t=1}^T Pθ(xt-1|xt)", + "content": "where the joint distribution of the model is the product of the learned conditionals: Pθ(x0, . . . , xT ) =Q(x T ) TY t=1 Pθ(xt−1|xt).(5)", + "attributes": [ + { + "attribute": "equation", + "value": "(5)" + }, + { + "attribute": "section", + "value": "joint distribution" + } + ] + }, + { + "type": "fact", + "insight": "Energy-Based Models re-cast the forward process in exponential form: Q(xt|xt-1) ∝ e^(-Ef_t-1(xt-1,xt))", + "content": "In many cases, it is straight- forward to re-cast the forward process in an exponential form, Q(xt|xt−1)∝e −Ef t−1(xt−1,xt),(6)", + "attributes": [ + { + "attribute": "method", + "value": "EBM re-casting" + }, + { + "attribute": "equation", + "value": "(6)" + }, + { + "attribute": "section", + "value": "EBM approach" + } + ] + }, + { + "type": "fact", + "insight": "DTMs generalize EBM approach by introducing latent variables {zt}, allowing independent scaling of model size/complexity from data dimension", + "content": "To maximally leverage probabilistic hardware for EBM sampling, DTMs generalize Eq. (7) by introducing latent variables{z t}: Pθ(xt−1|xt)∝ X zt−1 e−(Ef t−1(xt−1,xt)+Eθ t−1(xt−1,zt−1,θ)). (8) Introducing latent variables allows the size and complexity of the probabilistic model to be increased independently of the data dimension.", + "attributes": [ + { + "attribute": "method", + "value": "DTM generalization" + }, + { + "attribute": "equation", + "value": "(8)" + }, + { + "attribute": "innovation", + "value": "latent variables" + } + ] + }, + { + "type": "fact", + "insight": "DTMs have the property that exact reverse-process conditional approximation also learns marginal distribution at t-1: Q(xt-1) ∝ ∑_{zt-1} e^(-Eθ_t-1(xt-1,zt-1,θ))", + "content": "A convenient property of DTMs is that if the ap- proximation to the reverse-process conditional is exact (Pθ(xt−1|xt)→Q(x t−1|xt)), one also learns the marginal distribution att−1, Q(xt−1)∝ X zt−1 e−Eθ t−1(xt−1,zt−1,θ).(9)", + "attributes": [ + { + "attribute": "property", + "value": "marginal learning" + }, + { + "attribute": "equation", + "value": "(9)" + }, + { + "attribute": "condition", + "value": "exact reverse-process approximation" + } + ] + }, + { + "type": "opinion", + "insight": "Increasing T while holding EBM architecture constant increases expressive power and makes each sampling step easier, bypassing MET constraints", + "content": "As the number of steps in the forward process is increased, the effect of each noising step becomes smaller, meaning that Ef t−1 more tightly bindsx t tox t−1. This binding can simplify the distribution given in Eq. (7)... As illustrated in Fig. 3 (a), models of the form given in Eq. (7) reshape simple noise into an approximation of the data distribution. IncreasingTwhile holding the EBM architecture constant simultaneously increases the expressive power of the chain and makes each step easier to sample from, entirely bypassing the MET.", + "attributes": [ + { + "attribute": "benefit", + "value": "increased expressive power" + }, + { + "attribute": "advantage", + "value": "easier sampling" + }, + { + "attribute": "figure", + "value": "Fig. 3(a)" + } + ] + }, + { + "type": "comment", + "insight": "DTCA architecture tightly integrates DTMs into probabilistic hardware for highly efficient implementation, with each EBM implemented by distinct circuitry for input/output conditioning and latent sampling", + "content": "The Denoising Thermodynamic Computer Architec- ture (DTCA) tightly integrates DTMs into probabilistic hardware, allowing for the highly efficient implementa- (b)A sketch of how a chip based on the DTCA chains hard- ware EBMs to approximate the reverse process. Each EBM is implemented by distinct circuitry, parts of which are dedicated to receiving the inputs and conditionally sampling the outputs and latents.", + "attributes": [ + { + "attribute": "architecture", + "value": "DTCA" + }, + { + "attribute": "implementation", + "value": "probabilistic hardware" + }, + { + "attribute": "hardware_component", + "value": "distinct circuitry per EBM" + } + ] + }, + { + "type": "fact", + "insight": "The DTCA architecture uses constrained Energy-Based Models (EBMs) with sparse and local connectivity that can be implemented using massively parallel arrays of primitive circuitry performing Gibbs sampling.", + "content": "Practical implementations of the DTCA utilize natural-to-implement EBMs that exhibit sparse and local connectivity, as is typical in the literature [33]. This constraint allows sampling of the EBM to be performed by massively parallel arrays of primitive circuitry that implement Gibbs sampling.", + "attributes": [ + { + "attribute": "source", + "value": "DTCA architecture description" + }, + { + "attribute": "technical_approach", + "value": "hardware implementation" + }, + { + "attribute": "reference", + "value": "[33]" + } + ] + }, + { + "type": "fact", + "insight": "The reverse process transformation Ef t−1 can be implemented efficiently using pairwise interactions between variables in xt and xt−1, with no constraints on the form of Eθ t−1.", + "content": "A key feature of the DTCA is thatEf t−1 can be implemented efficiently using our constrained EBMs. Specifically, for both continuous and discrete diffusion,E f t−1 can be implemented using a single pairwise interaction between corresponding variables inxt andx t−1; see Ap- pendix A.1 and C.1 for details. This structure can be reflected in how the chip is laid out to implement these interactions without violating locality constraints. Critically, Eq. (8) places no constraints on the form ofE θ t−1. Therefore, we are free to use EBMs that our hardware implements especially efficiently.", + "attributes": [ + { + "attribute": "technical_approach", + "value": "algorithm efficiency" + }, + { + "attribute": "constraint_type", + "value": "pairwise interactions" + }, + { + "attribute": "reference", + "value": "Eq. (8)" + } + ] + }, + { + "type": "fact", + "insight": "The DTM performance was tested on Fashion-MNIST dataset using GPU simulation with FID metrics and energy consumption estimates, compared against conventional VAE on GPU.", + "content": "To understand the performance of a future hardware device, we developed a GPU simulator of the DTCA and used it to train a DTM on the Fashion-MNIST dataset. We measure the performance of the DTM using FID and utilize a physical model to estimate the energy required to generate new images. These numbers can be compared to conventional algorithm/hardware pairings, such as a VAE running on a GPU; these results are shown in Fig. 1.", + "attributes": [ + { + "attribute": "evaluation_method", + "value": "GPU simulation" + }, + { + "attribute": "dataset", + "value": "Fashion-MNIST" + }, + { + "attribute": "metrics", + "value": "FID, energy consumption" + }, + { + "attribute": "comparison_baselines", + "value": "VAE on GPU" + } + ] + }, + { + "type": "fact", + "insight": "Boltzmann machines with binary random variables were used as the simplest type of discrete-variable EBM, implementing energy functions with specific mathematical form E(x) = −β(Σi≠j xiJijxj + Σi hi xi).", + "content": "The DTM that produced the results shown in Fig. 1 used Boltzmann machine EBMs. Boltzmann machines, also known as Ising models in physics, use binary random variables and are the simplest type of discrete-variable EBM. Boltzmann machines implement energy functions of the form E(x) =−β 〈X i̸=j xiJijxj + X i=1 hixi 〉,(10) where eachx i ∈ {−1,1}.", + "attributes": [ + { + "attribute": "model_type", + "value": "Boltzmann machine" + }, + { + "attribute": "variable_type", + "value": "binary random variables" + }, + { + "attribute": "alternative_name", + "value": "Ising models" + }, + { + "attribute": "equation_reference", + "value": "Eq. (10)" + } + ] + }, + { + "type": "fact", + "insight": "The Gibbs sampling update rule for Boltzmann machines follows a sigmoidal probability function P(Xi[k+1] = +1|X[k] = x) = σ(2β(Σj≠i Jij xj + hi)), which can be implemented using simple circuitry with appropriately biased random bits.", + "content": "The Gibbs sampling update rule for sampling from the corresponding EBM is P(X i[k+ 1] = +1|X[k] =x) =σ 〈2β 〈X j̸=i Jij xj+hi 〉〉, (11) which can be evaluated simply using an appropriately biased source of random bits.", + "attributes": [ + { + "attribute": "algorithm", + "value": "Gibbs sampling" + }, + { + "attribute": "probability_function", + "value": "sigmoidal" + }, + { + "attribute": "implementation_medium", + "value": "random bits" + }, + { + "attribute": "equation_reference", + "value": "Eq. (11)" + } + ] + }, + { + "type": "fact", + "insight": "The Boltzmann machines were implemented as sparse, deep models with L×L grids (L=70), where each variable connects to several neighbors (typically 12), following bipartite connectivity patterns for parallel sampling.", + "content": "Specifically, the EBMs employed in this work were sparse, deep Boltzmann machines comprisingL×Lgrids of binary variables, whereL= 70was used in most cases. Eachvariablewasconnectedtoseveral(inmostcases, 12) of its neighbors following a simple pattern. At random, some of the variables were selected to represent the data xt−1, and the rest were assigned to the latent variables zt−1. Then, an extra node was connected to each data node to implement the coupling toxt.", + "attributes": [ + { + "attribute": "architecture", + "value": "sparse, deep Boltzmann machines" + }, + { + "attribute": "grid_size", + "value": "L×L, L=70" + }, + { + "attribute": "connectivity", + "value": "12 neighbors typically" + }, + { + "attribute": "variable_types", + "value": "data nodes, latent variables" + } + ] + }, + { + "type": "fact", + "insight": "The shot-noise dynamics of subthreshold transistors were used to build an RNG that is fast, energy-efficient, and small, with experimental results showing sigmoidal response to control voltage and approximately exponential autocorrelation decaying in ~100ns.", + "content": "To enable a near-term, large-scale realization of the DTCA, we leveraged the shot-noise dynamics of sub- threshold transistors [45] to build an RNG that is fast, energy-efficient, and small. Our all-transistor RNG is programmable and has the desired sigmoidal response to a control voltage, as shown by experimental measurements in Fig. 4 (a). The stochastic voltage signal output from the RNG has an approximately exponential autocorrelation function that decays in around100ns, as il- lustrated in Fig. 4 (b).", + "attributes": [ + { + "attribute": "hardware_component", + "value": "RNG" + }, + { + "attribute": "implementation_technology", + "value": "subthreshold transistors" + }, + { + "attribute": "properties", + "value": "fast, energy-efficient, small, programmable" + }, + { + "attribute": "response_characteristic", + "value": "sigmoidal" + }, + { + "attribute": "correlation_time", + "value": "~100ns" + } + ] + }, + { + "type": "comment", + "insight": "The modular nature of DTMs enables flexible hardware implementations including distinct physical circuitry per EBM, split across communicating chips, or reprogrammable hardware with different weights at different times.", + "content": "The modular nature of DTMs enables various hardware implementations. For example, each EBM in the chain can be implemented using distinct physical circuitry on the same chip, as shown in Fig. 3 (b). Alternatively, the various EBMs may be split across several communicating chips or implemented by the same hardware, reprogrammed with distinct sets of weights at different times.", + "attributes": [ + { + "attribute": "design_approach", + "value": "modular architecture" + }, + { + "attribute": "flexibility", + "value": "multiple implementation options" + }, + { + "attribute": "hardware_options", + "value": "distinct circuits, split chips, reprogrammable" + } + ] + }, + { + "type": "comment", + "insight": "A practical advantage of the all-transistor RNG design is that detailed foundry-provided models can be used to study manufacturing variations, enabling systematic design optimization.", + "content": "A practical advantage to our all-transistor RNG is that detailedandprovenfoundry-providedmodelscanbeused to study the effect of manufacturing variations on our", + "attributes": [ + { + "attribute": "design_advantage", + "value": "manufacturing variability analysis" + }, + { + "attribute": "model_availability", + "value": "foundry-provided models" + }, + { + "attribute": "optimization_approach", + "value": "systematic design" + } + ] + }, + { + "type": "fact", + "insight": "The sampling procedure uses block sampling of bipartite Boltzmann machines, where each color block can be sampled in parallel for K iterations (K≈1000, longer than mixing time) to draw samples from Eq. (7).", + "content": "Due to our chosen connectivity patterns, our Boltz-mann machines are bipartite (two-colorable). Since each color block can be sampled in parallel, a single itera- tion of Gibbs sampling corresponds to sampling the first colorblockconditionedonthesecondandthenviceversa. Starting from some random initialization, this block sampling procedure could then be repeated forKiterations (whereKis longer than the mixing time of the sampler, typicallyK≈1000) to draw samples from Eq. (7) for each step in the approximation to the reverse process.", + "attributes": [ + { + "attribute": "sampling_method", + "value": "block sampling" + }, + { + "attribute": "parallelization", + "value": "bipartite color blocks" + }, + { + "attribute": "iterations", + "value": "K≈1000" + }, + { + "attribute": "reference", + "value": "Eq. (7)" + } + ] + }, + { + "type": "fact", + "insight": "DTCA implementation uses sparse and locally connected EBMs that enable massively parallel Gibbs sampling via primitive circuitry", + "content": "Practical implementations of the DTCA utilize natural-to-implement EBMs that exhibit sparse and local connectivity, as is typical in the literature [33]. This constraint allows sampling of the EBM to be performed by massively parallel arrays of primitive circuitry that implement Gibbs sampling.", + "attributes": [ + { + "attribute": "source", + "value": "Literature [33]" + }, + { + "attribute": "implementation_type", + "value": "Hardware circuitry" + }, + { + "attribute": "sampling_method", + "value": "Gibbs sampling" + } + ] + }, + { + "type": "fact", + "insight": "Ef t−1 can be implemented efficiently using constrained EBMs for both continuous and discrete diffusion", + "content": "A key feature of the DTCA is thatEf t−1 can be implemented efficiently using our constrained EBMs. Specifically, for both continuous and discrete diffusion,E f t−1 can be implemented using a single pairwise interaction between corresponding variables inxt andx t−1.", + "attributes": [ + { + "attribute": "feature", + "value": "DTCA efficiency" + }, + { + "attribute": "diffusion_types", + "value": "Continuous and discrete" + }, + { + "attribute": "implementation", + "value": "Pairwise interaction" + } + ] + }, + { + "type": "fact", + "insight": "EBMs can be scaled arbitrarily by combining hardware latent-variable EBMs into software-defined graphical models", + "content": "At the lowest level, this corresponds to high-dimensional, regularly structured latent variable EBM. If more powerful models are desired, these hardware latent-variable EBMs can be arbitrarily scaled by combining them into software-defined graphical models.", + "attributes": [ + { + "attribute": "scalability", + "value": "Arbitrary scaling" + }, + { + "attribute": "model_type", + "value": "Graphical models" + } + ] + }, + { + "type": "fact", + "insight": "DTM hardware implementations offer modular design options including distinct circuitry per EBM, split across chips, or reprogrammable hardware", + "content": "The modular nature of DTMs enables various hardware implementations. For example, each EBM in the chain can be implemented using distinct physical circuitry on the same chip, as shown in Fig. 3 (b). Alternatively, the various EBMs may be split across several communicating chips or implemented by the same hardware, reprogrammed with distinct sets of weights at different times.", + "attributes": [ + { + "attribute": "modularity", + "value": "Various implementations" + }, + { + "attribute": "reference", + "value": "Fig. 3 (b)" + } + ] + }, + { + "type": "fact", + "insight": "GPU simulator developed for DTCA achieved FID-validated performance on Fashion-MNIST dataset with energy efficiency compared to VAE on GPU", + "content": "To understand the performance of a future hardware device, we developed a GPU simulator of the DTCA and used it to train a DTM on the Fashion-MNIST dataset. We measure the performance of the DTM using FID and utilize a physical model to estimate the energy required to generate new images. These numbers can be compared to conventional algorithm/hardware pairings, such as a VAE running on a GPU; these results are shown in Fig. 1.", + "attributes": [ + { + "attribute": "dataset", + "value": "Fashion-MNIST" + }, + { + "attribute": "performance_metric", + "value": "FID" + }, + { + "attribute": "reference", + "value": "Fig. 1" + } + ] + }, + { + "type": "fact", + "insight": "Boltzmann machines serve as hardware-efficient EBMs due to simple Gibbs sampling update rules", + "content": "Boltzmann machines are hardware efficient because the Gibbs sampling update rule required to sample from them is simple. Boltzmann machines implement energy functions of the form E(x) =−β⟨∑i̸=j xiJijxj + ∑i=1 hixi⟩,(10), where eachx i ∈ {−1,1}.", + "attributes": [ + { + "attribute": "efficiency_reason", + "value": "Simple Gibbs sampling" + }, + { + "attribute": "variable_type", + "value": "Binary" + }, + { + "attribute": "reference", + "value": "Eq. (10)" + } + ] + }, + { + "type": "fact", + "insight": "Boltzmann machine hardware implementation uses regular grid of Bernoulli sampling circuits with sigmoidal bias control", + "content": "Implementing our proposed hardware architecture using Boltzmann machines is particularly simple. A device will consist of a regular grid of Bernoulli sampling circuits, where each sampling circuit implements the Gibbs sampling update for a single variablex i. The bias of the sampling circuits (probability that it produces 1 as opposed to−1) is constrained to be a sigmoidal function of an input voltage, allowing the conditional update given in Eq. (11) to be implemented using a simple circuit that adds currents such as a resistor network.", + "attributes": [ + { + "attribute": "circuit_type", + "value": "Bernoulli sampling circuits" + }, + { + "attribute": "bias_control", + "value": "Sigmoidal function" + }, + { + "attribute": "reference", + "value": "Eq. (11)" + } + ] + }, + { + "type": "fact", + "insight": "Boltzmann machines use 70×70 grids with 12 neighbor connections in bipartite configuration", + "content": "Due to our chosen connectivity patterns, our Boltzmann machines are bipartite (two-colorable). Since each color block can be sampled in parallel, a single iteration of Gibbs sampling corresponds to sampling the first colorblockconditiononthesecondandthenviceversa. Starting from some random initialization, this block sampling procedure could then be repeated forKiterations (whereKis longer than the mixing time of the sampler, typicallyK≈1000) to draw samples from Eq. (7) for each step in the approximation to the reverse process.", + "attributes": [ + { + "attribute": "grid_size", + "value": "70×70" + }, + { + "attribute": "connections", + "value": "12 neighbors" + }, + { + "attribute": "iterations", + "value": "K≈1000" + } + ] + }, + { + "type": "fact", + "insight": "Subthreshold transistor shot-noise dynamics enabled fast, energy-efficient programmable RNG with 100ns autocorrelation decay", + "content": "To enable a near-term, large-scale realization of the DTCA, we leveraged the shot-noise dynamics of subthreshold transistors [45] to build an RNG that is fast, energy-efficient, and small. Our all-transistor RNG is programmable and has the desired sigmoidal response to a control voltage, as shown by experimental measurements in Fig. 4 (a). The stochastic voltage signal output from the RNG has an approximately exponential autocorrelation function that decays in around100ns, as illustrated in Fig. 4 (b).", + "attributes": [ + { + "attribute": "technology", + "value": "Subthreshold transistors [45]" + }, + { + "attribute": "performance", + "value": "100ns decay" + }, + { + "attribute": "reference", + "value": "Fig. 4 (a), (b)" + } + ] + }, + { + "type": "opinion", + "insight": "DTCA hardware efficiency is a key advantage due to unconstrained Eθ t−1 allowing selection of hardware-optimized EBMs", + "content": "Critically, Eq. (8) places no constraints on the form ofE θ t−1. Therefore, we are free to use EBMs that our hardware implements especially efficiently.", + "attributes": [ + { + "attribute": "advantage", + "value": "Hardware efficiency" + }, + { + "attribute": "constraint", + "value": "None on Eθ t−1" + } + ] + }, + { + "type": "opinion", + "insight": "Near-term DTCA realization requires leveraging transistor shot-noise dynamics for practical RNG implementation", + "content": "To enable a near-term, large-scale realization of the DTCA, we leveraged the shot-noise dynamics of subthreshold transistors [45] to build an RNG that is fast, energy-efficient, and small.", + "attributes": [ + { + "attribute": "strategy", + "value": "Near-term realization" + }, + { + "attribute": "component", + "value": "RNG implementation" + } + ] + }, + { + "type": "comment", + "insight": "DTCA architecture references multiple appendices (B, C, A.1, C.1, D.1, C, J) for theoretical discussion and implementation details", + "content": "Refer to Appendices B and C for a further theoretical discussion of the hardware architecture. See Appendix D.1. Appendix C provides further details on the Boltzmann machine architecture. Appendix J provides further details about our RNG.", + "attributes": [ + { + "attribute": "documentation", + "value": "Multiple appendices" + }, + { + "attribute": "topics", + "value": "Theoretical discussion, architecture details" + } + ] + }, + { + "type": "comment", + "insight": "DTCA performance evaluation includes FID metrics and energy consumption comparisons against VAE baseline", + "content": "We measure the performance of the DTM using FID and utilize a physical model to estimate the energy required to generate new images. These numbers can be compared to conventional algorithm/hardware pairings, such as a VAE running on a GPU; these results are shown in Fig. 1.", + "attributes": [ + { + "attribute": "evaluation_metrics", + "value": "FID, Energy consumption" + }, + { + "attribute": "baseline", + "value": "VAE on GPU" + }, + { + "attribute": "reference", + "value": "Fig. 1" + } + ] + }, + { + "type": "fact", + "insight": "The document describes a programmable random number generator (RNG) with operating characteristics that can be controlled by varying an input voltage, with the probability of high state output following a sigmoid function relationship.", + "content": "FIG. 4.A programmable source of random bits. (a)A laboratory measurement of the operating characteristic of our RNG. The probability of the output voltage signal being in the high state (x= 1) can be programmed by varying an input voltage. The relationship betweenP(x= 1)and the input voltage is well-approximated by a sigmoid function.", + "attributes": [ + { + "attribute": "source", + "value": "Fig. 4(a)" + }, + { + "attribute": "component", + "value": "RNG operating characteristics" + } + ] + }, + { + "type": "fact", + "insight": "The RNG's autocorrelation function shows exponential decay with a time constant τ0 ≈ 100ns at the unbiased point (P(x=1) = 0.5), indicating good random behavior.", + "content": "(b)The autocorrelation function of the RNG at the unbiased point (P(x= 1) = 0.5). The decay is approximately exponential with the rateτ0 ≈100ns.", + "attributes": [ + { + "attribute": "source", + "value": "Fig. 4(b)" + }, + { + "attribute": "measurement", + "value": "autocorrelation decay" + }, + { + "attribute": "value", + "value": "τ0 ≈ 100ns" + } + ] + }, + { + "type": "fact", + "insight": "Manufacturing variation studies show that the RNG works reliably across different process corners, with the 'slow NMOS, fast PMOS' case being the worst performer due to design asymmetry.", + "content": "(c)Estimating the effect of manufacturing variation on RNG performance. Each point in the plot represents the results of a simulation of an RNG circuit with transistor parameters sampled according to a procedure defined by the manufacturer's PDK. Each color represents a different process corner, each for which∼200realizations of the RNG were simulated. The \"typical\" corner represents a balanced case, whereas the other two are asymmetric corners where the two types of transistors (NMOS and PMOS) are skewed in opposite directions. The slow NMOS and fast PMOS case is worst performing for us due to an asymmetry in our design.", + "attributes": [ + { + "attribute": "source", + "value": "Fig. 4(c)" + }, + { + "attribute": "analysis_type", + "value": "manufacturing variation" + }, + { + "attribute": "finding", + "value": "Reliable across process corners" + } + ] + }, + { + "type": "fact", + "insight": "The energy consumption of the probabilistic computer is modeled using a physical model of an all-transistor Boltzmann machine Gibbs sampler, with energy contributions from RNG, bias, clock, and communication components.", + "content": "The energy estimates given in Fig. 1 for the probabilistic computer were constructed using a physical model of an all-transistor Boltzmann machine Gibbs sampler. The dominant contributions to this model are captured by the formula E=T KmixL2Ecell,(12) Ecell =E rng +E bias +E clock +E comm,(13) whereE rng comes from the data in Fig. 4 (c).", + "attributes": [ + { + "attribute": "model_type", + "value": "Boltzmann machine Gibbs sampler" + }, + { + "attribute": "components", + "value": "RNG, bias, clock, communication" + }, + { + "attribute": "equation", + "value": "Ecell = E_rng + E_bias + E_clock + E_comm" + } + ] + }, + { + "type": "fact", + "insight": "The estimated cell energy consumption Ecell ≈ 2fJ is derived from a physical model using the same transistor process as the RNG with reasonable parameter selections.", + "content": "Generally, given the same transistor process we used for our RNG and some reasonable selections for other free parameters of the model, we can estimate Ecell ≈2fJ. See Appendix D for an exhaustive derivation of this model.", + "attributes": [ + { + "attribute": "value", + "value": "Ecell ≈ 2fJ" + }, + { + "attribute": "derivation", + "value": "Physical model" + }, + { + "attribute": "source", + "value": "Appendix D" + } + ] + }, + { + "type": "fact", + "insight": "Energy consumption for GPU implementation is estimated by computing total floating-point operations (FLOPs) required and dividing by manufacturer's FLOP/joule specification.", + "content": "We use a simple model for the energy consumption of the GPU that underestimates the actual values. We compute the total number of floating-point operations (FLOPs) required to generate a sample from the trained model and divide that by the FLOP/joule specification given by the manufacturer.", + "attributes": [ + { + "attribute": "method", + "value": "FLOPs / manufacturer specification" + }, + { + "attribute": "component", + "value": "GPU energy estimation" + }, + { + "attribute": "reference", + "value": "Appendix E" + } + ] + }, + { + "type": "fact", + "insight": "Energy-based models (EBMs) are trained using Monte-Carlo estimators for gradient computation, with terms computed independently for each time step t.", + "content": "The EBMs used in the experiments presented in Fig. 1 were trained by applying the standard Monte-Carlo estimator for the gradients of EBMs [61] to Eq. (4), which yields ∇θLDN (θ)= TX t=1 EQ(xt−1,xt) [EPθ(zt−1|xt−1,xt) [∇θEm t−1 ] −EPθ(xt−1,zt−1|xt) [∇θEm t−1 ] ] . (14) Notably, each term in the sum overtcan be computed independently.", + "attributes": [ + { + "attribute": "method", + "value": "Monte-Carlo gradient estimation" + }, + { + "attribute": "equation", + "value": "Eq. (14)" + }, + { + "attribute": "characteristic", + "value": "Independent term computation" + } + ] + }, + { + "type": "fact", + "insight": "DTMs allow EBMs to have finite and short mixing times, enabling nearly unbiased gradient estimates through sufficient sampling iterations, unlike MEBMs which typically have long mixing times making unbiased gradient estimates impossible in most cases.", + "content": "It should be noted that the DTCA allows our EBMs to have finite and short mixing times, which enables sufficient sampling iterations to be used to achieve nearly unbiased estimates of the gradient. Unbiased gradient estimates are not possible for MEBMs in most cases due to their long mixing times [62].", + "attributes": [ + { + "attribute": "source", + "value": "text" + }, + { + "attribute": "reference", + "value": "[62]" + } + ] + }, + { + "type": "fact", + "insight": "DTMs significantly improve training stability compared to MEBMs, and when complemented with ACP (Annealed Control Process), completely stabilize the training process.", + "content": "DTMs alleviate the training instability that is fundamental to MEBMs... An example of the training dynamics for several different types of models is shown in Fig. 5 (b)... Complementing DTMs with the ACP completely stabilizes training.", + "attributes": [ + { + "attribute": "source", + "value": "text" + }, + { + "attribute": "reference", + "value": "Fig. 5(b)" + }, + { + "attribute": "method", + "value": "DTM + ACP" + } + ] + }, + { + "type": "comment", + "insight": "MEBM training becomes unstable as the model becomes complex and multimodal during training, causing samples to deviate from equilibrium and gradients to lose meaningful direction.", + "content": "However, as these gradients are followed, the MEBM is reshaped according to the data distribution and begins to become complex and multimodal. This induced multimodality greatly increases the sampling complexity of the distribution, causing samples to deviate from equilibrium. Gradients computed using non-equilibrium samples do not necessarily point in a meaningful direction, which can halt or, in some cases, even reverse the training process.", + "attributes": [ + { + "attribute": "source", + "value": "text" + }, + { + "attribute": "problem", + "value": "training instability" + } + ] + }, + { + "type": "fact", + "insight": "Training stability can be measured using normalized autocorrelation (ryy[k]) where values close to 1 indicate far-from-equilibrium samples and low-quality gradients, while values close to 0 indicate samples near equilibrium and high-quality gradient estimates.", + "content": "The lower plot in Fig. 5 (b) shows the autocorrelation at a delay equal to the total number of sampling iterations used to estimate the gradients during training. Generally, if r_yy is close to 1, gradients were estimated using far-from-equilibrium samples and were likely of low quality. If it is close to zero, the samples should be close to equilibrium and produce high-quality gradient estimates.", + "attributes": [ + { + "attribute": "source", + "value": "text" + }, + { + "attribute": "metric", + "value": "normalized autocorrelation r_yy[k]" + }, + { + "attribute": "reference", + "value": "Eq. (15), (16)" + } + ] + }, + { + "type": "comment", + "insight": "Denoising models stabilize training by implementing simpler transformations per layer, reducing the complexity of the distribution the model must learn and making it easier to sample from.", + "content": "Denoising alone significantly stabilizes training. Because the transformation carried out by each layer is simpler, the distribution that the model must learn is less complex and, therefore, easier to sample from.", + "attributes": [ + { + "attribute": "source", + "value": "text" + }, + { + "attribute": "method", + "value": "denoising" + } + ] + }, + { + "type": "fact", + "insight": "EBM performance scales with complexity - layers with more connectivity and longer allowed mixing times can utilize more latent variables and achieve higher performance, as demonstrated by varying grid size L in Fashion-MNIST experiments.", + "content": "The effect of scaling EBM complexity on DTM performance. The grid size L was modified to change the number of latent variables compared to the (fixed) number of data variables. Generally, EBM layers with more connectivity and longer allowed mixing times can utilize more latent variables and, therefore, achieve higher performance.", + "attributes": [ + { + "attribute": "source", + "value": "text" + }, + { + "attribute": "reference", + "value": "Fig. 5(c)" + }, + { + "attribute": "dataset", + "value": "Fashion-MNIST" + } + ] + }, + { + "type": "fact", + "insight": "DTM training becomes unstable due to complex energy landscape development among latent variables", + "content": "As training progresses, the DTM eventually becomes unstable, which can be attributed to the development of a complex energy landscape among the latent variables.", + "attributes": [ + { + "attribute": "section", + "value": "Training Stability" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Total correlation penalty added to loss function to penalize poorly mixing models", + "content": "We add a term to the loss function that nudges the optimization towards a distribution that is easy to sample from", + "attributes": [ + { + "attribute": "section", + "value": "Training Procedure" + }, + { + "attribute": "method", + "value": "Total Correlation Penalty" + } + ] + }, + { + "type": "fact", + "insight": "Total loss function combines DN loss and total correlation penalty", + "content": "The total loss function is the sum of Eq. (4) and this total correlation penalty: L=L DN + Σ_{t=1}^{T} λtLTC t", + "attributes": [ + { + "attribute": "section", + "value": "Loss Function" + }, + { + "attribute": "equation", + "value": "(18)" + } + ] + }, + { + "type": "fact", + "insight": "Adaptive Correlation Penalty (ACP) provides closed-loop control of correlation penalty strengths", + "content": "We use an Adaptive Correlation Penalty (ACP) to set the λt as large as necessary to keep sampling tractable for each layer", + "attributes": [ + { + "attribute": "section", + "value": "Adaptive Control" + }, + { + "attribute": "method", + "value": "ACP" + } + ] + }, + { + "type": "fact", + "insight": "Increasing DTM depth from 2 to 8 substantially improves image generation quality", + "content": "As shown in Fig. 1, increasing the depth of the DTM from 2 to 8 substantially improves the quality of generated images", + "attributes": [ + { + "attribute": "section", + "value": "Scaling Analysis" + }, + { + "attribute": "figure", + "value": "Fig. 1" + }, + { + "attribute": "improvement", + "value": "substantial" + } + ] + }, + { + "type": "fact", + "insight": "Larger values of K are required to support wider models with constant connectivity", + "content": "which demonstrates that larger values of K are required to support wider models holding connectivity constant", + "attributes": [ + { + "attribute": "section", + "value": "Scaling Constraints" + }, + { + "attribute": "parameter", + "value": "K" + } + ] + }, + { + "type": "opinion", + "insight": "Probabilistic ML hardware should be scaled as part of hybrid systems rather than in isolation", + "content": "we hypothesize that the correct way to scale probabilistic machine learning hardware systems is not in isolation but rather as a component in a larger hybrid thermodynamic-deterministic machine learning (HTDML) system", + "attributes": [ + { + "attribute": "section", + "value": "Conclusion" + }, + { + "attribute": "hypothesis", + "value": "HTDML scaling approach" + } + ] + }, + { + "type": "opinion", + "insight": "Hardware-efficient EBM topology cannot be scaled in isolation to model arbitrarily complex datasets", + "content": "It would be naive to expect that a hardware-efficient EBM topology can be scaled in isolation to model arbitrarily complex datasets", + "attributes": [ + { + "attribute": "section", + "value": "Scaling Limitations" + }, + { + "attribute": "confidence", + "value": "qualified" + } + ] + }, + { + "type": "opinion", + "insight": "Deterministic processors are sometimes better tools for specific ML tasks", + "content": "A hybrid approach is sensible because there is no a priori reason to believe that a probabilistic computer should handle every part of a machine learning problem, and sometimes a deterministic processor is likely a better tool for the job", + "attributes": [ + { + "attribute": "section", + "value": "Hybrid Approach Rationale" + }, + { + "attribute": "rationale", + "value": "task-specific suitability" + } + ] + }, + { + "type": "comment", + "insight": "Training dynamics show monotonic quality improvement with small autocorrelations under closed-loop control", + "content": "Model quality increases monotonically, and the autocorrelation stays small throughout training. This closed-loop control of the correlation penalty was employed during the training of most models used to produce the results in this article", + "attributes": [ + { + "attribute": "section", + "value": "Training Dynamics" + }, + { + "attribute": "policy", + "value": "closed-loop control" + }, + { + "attribute": "figure", + "value": "Fig. 5 (b)" + } + ] + }, + { + "type": "comment", + "insight": "HTDML energy landscape decomposes into deterministic and probabilistic components", + "content": "Mathematically, the landscape of HTDML may be summarized as Etot(S, D, p) =Edet(S, D, p) +Eprob(S, D, p)", + "attributes": [ + { + "attribute": "section", + "value": "HTDML Formulation" + }, + { + "attribute": "equation", + "value": "(19)" + }, + { + "attribute": "components", + "value": "deterministic + probabilistic" + } + ] + }, + { + "type": "fact", + "insight": "A DTM trained to generate CIFAR-10 images achieves performance parity with a traditional GAN using approximately 10× smaller deterministic neural network", + "content": "The DTM is trained to generate CIFAR-10 images and achieves performance parity with a traditional GAN using a∼10×smaller deterministic neural network.", + "attributes": [ + { + "attribute": "source", + "value": "Figure 6 description" + }, + { + "attribute": "dataset", + "value": "CIFAR-10" + }, + { + "attribute": "comparison", + "value": "DTM vs traditional GAN" + } + ] + }, + { + "type": "fact", + "insight": "Binarization is not viable as a general approach for embedding data into hardware EBMs", + "content": "Indeed, binarization is not viable in general, and embedding into richer types of variables (such as categorical) at the probabilistic hardware level is not particularly efficient or principled.", + "attributes": [ + { + "attribute": "source", + "value": "Technical analysis section" + }, + { + "attribute": "method", + "value": "binarization critique" + }, + { + "attribute": "scope", + "value": "general applicability" + } + ] + }, + { + "type": "comment", + "insight": "Current embedding methods using autoencoders and DTMs are not jointly trained, which may result in suboptimal embedding due to DTM's limited connectivity", + "content": "One major flaw with our method is that the autoencoder and DTM are not jointly trained, which means that the embedding learned by the autoencoder may not be well-suited to the way information can flow in the DTM, given its limited connectivity.", + "attributes": [ + { + "attribute": "source", + "value": "Analysis of embedding method" + }, + { + "attribute": "limitation", + "value": "joint training not implemented" + }, + { + "attribute": "hardware_constraint", + "value": "DTM limited connectivity" + } + ] + }, + { + "type": "fact", + "insight": "A 6×6µm chip could potentially fit approximately 10^6 sampling cells based on current RNG size estimates", + "content": "Based on the size of our RNG, it can be estimated that∼10 6 sampling cells could be fit into a6×6µm chip (see Appendix J).", + "attributes": [ + { + "attribute": "source", + "value": "Scalability analysis" + }, + { + "attribute": "chip_size", + "value": "6×6µm" + }, + { + "attribute": "capacity", + "value": "~10^6 sampling cells" + } + ] + }, + { + "type": "comment", + "insight": "There exists a significant gap between current model sizes and potential hardware capabilities, with the largest DTM using only around 50,000 cells compared to potential 10^6 capacity", + "content": "In contrast, the largest DTM shown in Fig. 1 would use only around 50,000 cells.", + "attributes": [ + { + "attribute": "source", + "value": "Scalability comparison" + }, + { + "attribute": "current_model_size", + "value": "~50,000 cells" + }, + { + "attribute": "potential_capacity", + "value": "~10^6 cells" + } + ] + }, + { + "type": "opinion", + "insight": "Optimal solutions in HTDML will likely be found between deterministic and probabilistic extremes where subsystem contributions are nearly balanced", + "content": "Like many engineered systems, optimal solutions will be found somewhere in the middle, where the contributions from the various subsystems are nearly balanced [65–67].", + "attributes": [ + { + "attribute": "source", + "value": "System design philosophy" + }, + { + "attribute": "approach", + "value": "balanced subsystem design" + }, + { + "attribute": "reference", + "value": "[65-67]" + } + ] + }, + { + "type": "fact", + "insight": "GPU simulation of hardware EBMs is inefficient due to sparse data structures not matching regular tensor data types", + "content": "One difficulty with HTDML research is that simulating large hardware EBMs on GPUs can be a challenging task. GPUs run these EBMs much less efficiently than probabilistic computers and the sparse data structures that naturally arise when working with hardware EBMs do not mesh well with regular tensor data types.", + "attributes": [ + { + "attribute": "source", + "value": "Research challenges" + }, + { + "attribute": "platform", + "value": "GPU simulation" + }, + { + "attribute": "issue", + "value": "sparse data structures vs regular tensors" + } + ] + }, + { + "type": "fact", + "insight": "A JAX-based software library with XLA acceleration has been developed for simulating hardware EBMs", + "content": "We have both short and long-term solutions to these challenges. To address these challenges in the short term, we have open-sourced a software library [69] that enables XLA-accelerated [70] simulation of hardware EBMs. This library is written in JAX [71] and automates the complex slicing operations that enable hardware EBM sampling.", + "attributes": [ + { + "attribute": "source", + "value": "Software solution" + }, + { + "attribute": "technology", + "value": "JAX with XLA acceleration" + }, + { + "attribute": "availability", + "value": "open-sourced" + } + ] + }, + { + "type": "comment", + "insight": "Page 10 contains a comprehensive bibliography section with 40 references spanning 2006-2025, covering topics in AI, machine learning, quantum computing, and related scientific fields.", + "content": "10\n[1] A. A. Chien, Commun. ACM66, 5 (2023).\n[2] D. D. Stine,The Manhattan Project, the Apollo Program,\nand Federal Energy Technology R&D Programs: A Com-\nparative Analysis, Report RL34645 (Congressional Re-\nsearch Service, Washington, D.C., 2009).\n[3] J. Aljbour, T. Wilson, and P. Patel, EPRI White Paper\nno. 3002028905 (2024).\n[4] Y. Li, D. Choi, J. Chung, N. Kushman, J. Schrittwieser,\nR. Leblond, T. Eccles, J. Keeling, F. Gimeno, A. D.\nLago, T. Hubert, P. Choy, C. de Masson d'Autume,\nI. Babuschkin, X. Chen, P.-S. Huang, J. Welbl, S. Gowal,\nA. Cherepanov, J. Molloy, D. J. Mankowitz, E. S. Rob-\nson, P. Kohli, N. de Freitas, K. Kavukcuoglu, and\nO. Vinyals, Science378, 1092 (2022).\n[5] D. M. Katz, M. J. Bommarito, S. Gao, and P. Arredondo,\nPhilos. Trans. R. Soc. A382, 20230254 (2024).\n[6] H. Nori, N. King, S. M. McKinney, D. Carignan, and\nE. Horvitz, arXiv [cs.CL] (2023).\n[7] S. Noy and W. Zhang, Science381, 187 (2023).\n[8] E. Brynjolfsson, D. Li, and L. Raymond, Q. J. Econ.\n10.1093/qje/qjae044 (2025).\n[9] S. Peng, E. Kalliamvakou, P. Cihon, and M. Demirer,\narXiv [cs.SE] (2023).\n[10] A. Bick, A. Blandin, and D. J. Deming, The rapid adop-\ntion of generative ai, Tech. Rep. (National Bureau of Eco-\nnomic Research, 2024).\n[11] A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit,\nL. Jones, A. N. Gomez, L. u. Kaiser, and I. Polosukhin,\ninAdvances in Neural Information Processing Systems,\nVol. 30, edited by I. Guyon, U. V. Luxburg, S. Bengio,\nH. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett\n(Curran Associates, Inc., 2017).\n[12] A. Coates, B. Huval, T. Wang, D. J. Wu, A. Y. Ng,\nand B. Catanzaro, inProceedings of the 30th Interna-\ntional Conference on International Conference on Ma-\nchine Learning - Volume 28, ICML'13 (JMLR.org, 2013)\np. III–1337–III–1345.\n[13] K. Chellapilla, S. Puri, and P. Simard, inTenth Inter-\nnational Workshop on Frontiers in Handwriting Recogni-\ntion, edited by G. Lorette, Université de Rennes 1 (Su-\nvisoft, La Baule (France), 2006).\n[14] H. Xiao, K. Rasul, and R. Vollgraf, arXiv [cs.LG] (2017).\n[15] M. Heusel, H. Ramsauer, T. Unterthiner, B. Nessler,\nand S. Hochreiter, inAdvances in Neural Information Pro-\ncessing Systems, Vol. 30, edited by I. Guyon, U. V.\nLuxburg, S. Bengio, H. Wallach, R. Fergus, S. Vish-\nwanathan, and R. Garnett (Curran Associates, Inc.,\n2017).\n[16] D.P.KingmaandM.Welling,Auto-Encoding Variational\nBayes(2022).\n[17] I. J. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu,\nD. Warde-Farley, S. Ozair, A. Courville, and Y. Ben-\nbio, inAdvances in Neural Information Processing Sys-\ntems, Vol. 27, edited by Z. Ghahramani, M. Welling,\nC. Cortes, N. Lawrence, and K. Weinberger (Curran As-\nsociates, Inc., 2014).\n[18] J. Sohl-Dickstein, E. Weiss, N. Maheswaranathan, and\nS. Ganguli, inProceedings of the 32nd International Con-\nference on Machine Learning, Proceedings of Machine\nLearning Research, Vol. 37, edited by F. Bach and D. Blei\n(PMLR, Lille, France, 2015) pp. 2256–2265.\n[19] S. Hooker, Commun. ACM64, 58–65 (2021).\n[20] S. Ambrogio, P. Narayanan, A. Okazaki, A. Fasoli,\nC. Mackin, K. Hosokawa, A. Nomura, T. Yasuda,\nA. Chen, A. Friz,et al., Nature620, 768 (2023).\n[21] S. Bandyopadhyay, A. Sludds, S. Krastanov, R. Hamerly,\nN. Harris, D. Bunandar, M. Streshinsky, M. Hochberg,\nand D. Englund, Nat. Photon.18, 1335 (2024).\n[22] H. A. Gonzalez, J. Huang, F. Kelber, K. K. Nazeer,\nT.Langer, C.Liu, M.Lohrmann, A.Rostami, M.Schone,\nB. Vogginger,et al., arXiv [cs.ET] (2024).\n[23] S. B. Shrestha, J. Timcheck, P. Frady, L. Campos-\nMacias, and M. Davies, inICASSP 2024 - 2024 IEEE\nInternational Conference on Acoustics, Speech and Sig-\nnal Processing (ICASSP)(2024) pp. 13481–13485.\n[24] Y. Sun, N. B. Agostini, S. Dong, and D. Kaeli, arXiv\n[cs.DC] (2019).\n[25] Y. Song and S. Ermon, inAdvances in Neural Informa-\ntion Processing Systems, Vol. 32, edited by H. Wallach,\nH. Larochelle, A. Beygelzimer, F. d'Alché-Buc, E. Fox,\nand R. Garnett (Curran Associates, Inc., 2019).\n[26] M. Janner, Y. Du, J. Tenenbaum, and S. Levine, inInter-\nnational Conference on Machine Learning(PMLR, 2022)\npp. 9902–9915.\n[27] N. S. Singh, K. Kobayashi, Q. Cao, K. Selcuk, T. Hu,\nS. Niazi, N. A. Aadit, S. Kanai, H. Ohno, S. Fukami,\net al., Nat. Commun.15, 2685 (2024).\n[28] C. Pratt, K. Ray, and J. Crutchfield,Dynamical Com-\nputing on the Nanoscale: Superconducting Circuits for\nThermodynamically-Efficient Classical Information Pro-\ncessing(2023).\n[29] G. Wimsatt, O.-P. Saira, A. B. Boyd, M. H. Matheny,\nS. Han, M. L. Roukes, and J. P. Crutchfield, Phys. Rev.\nRes.3, 033115 (2021).\n[30] S. H. Adachi and M. P. Henderson,Application of Quan-\ntum Annealing to Training of Deep Neural Networks\n(2015).\n[31] B. Sutton, K. Y. Camsari, B. Behin-Aein, and S. Datta,\nSci. Rep.7, 44370 (2017).\n[32] R. Faria, K. Y. Camsari, and S. Datta, IEEE Magn. Lett.\n8, 1 (2017).\n[33] S.Niazi, S.Chowdhury, N.A.Aadit, M.Mohseni, Y.Qin,\nand K. Y. Camsari, Nat. Electron.7, 610 (2024).\n[34] W. A. Borders, A. Z. Pervaiz, S. Fukami, K. Y. Camsari,\nH. Ohno, and S. Datta, Nature573, 390 (2019).\n[35] N. S. Singh, K. Kobayashi, Q. Cao, K. Selcuk, T. Hu,\nS. Niazi, N. A. Aadit, S. Kanai, H. Ohno, S. Fukami,\net al., Nat. Commun.15, 2685 (2024).\n[36] M. M. H. Sajeeb, N. A. Aadit, S. Chowdhury, T. Wu,\nC. Smith, D. Chinmay, A. Raut, K. Y. Camsari, C. Dela-\ncour, and T. Srimani, Phys. Rev. Appl.24, 014005\n(2025).\n[37] T. Conte, E. DeBenedictis, N. Ganesh, T. Hylton,\nJ. P. Strachan, R. S. Williams, A. Alemi, L. Altenberg,\nG. Crooks, J. Crutchfield,et al., arXiv [cs.CY] (2019).\n[38] Y. Du and I. Mordatch, inAdvances in Neural Informa-\ntion Processing Systems, Vol. 32, edited by H. Wallach,\nH. Larochelle, A. Beygelzimer, F. d'Alché-Buc, E. Fox,\nand R. Garnett (Curran Associates, Inc., 2019).\n[39] W. Lee, H. Kim, H. Jung, Y. Choi, J. Jeon, and C. Kim,\nSci. Rep.15, 8018 (2025).\n[40] M. Horodynski, C. Roques-Carmes, Y. Salamin, S. Choi,", + "attributes": [ + { + "attribute": "page_number", + "value": "10" + }, + { + "attribute": "section", + "value": "References/Bibliography" + }, + { + "attribute": "total_entries", + "value": "40" + }, + { + "attribute": "date_range", + "value": "2006-2025" + } + ] + }, + { + "type": "fact", + "insight": "The bibliography includes recent cutting-edge AI research from 2022-2025, including generative AI adoption studies, neural network research, and quantum computing applications.", + "content": "[4] Y. Li, D. Choi, J. Chung, N. Kushman, J. Schrittwieser,\nR. Leblond, T. Eccles, J. Keeling, F. Gimeno, A. D.\nLago, T. Hubert, P. Choy, C. de Masson d'Autume,\nI. Babuschkin, X. Chen, P.-S. Huang, J. Welbl, S. Gowal,\nA. Cherepanov, J. Molloy, D. J. Mankowitz, E. S. Rob-\nson, P. Kohli, N. de Freitas, K. Kavukcuoglu, and\nO. Vinyals, Science378, 1092 (2022).\n[5] D. M. Katz, M. J. Bommarito, S. Gao, and P. Arredondo,\nPhilos. Trans. R. Soc. A382, 20230254 (2024).\n[6] H. Nori, N. King, S. M. McKinney, D. Carignan, and\nE. Horvitz, arXiv [cs.CL] (2023).\n[7] S. Noy and W. Zhang, Science381, 187 (2023).\n[8] E. Brynjolfsson, D. Li, and L. Raymond, Q. J. Econ.\n10.1093/qje/qjae044 (2025).\n[10] A. Bick, A. Blandin, and D. J. Deming, The rapid adop-\ntion of generative ai, Tech. Rep. (National Bureau of Eco-\nnomic Research, 2024).", + "attributes": [ + { + "attribute": "page_number", + "value": "10" + }, + { + "attribute": "focus_area", + "value": "Recent AI Research (2022-2025)" + }, + { + "attribute": "publication_types", + "value": "Science, arXiv, Economic Research" + } + ] + }, + { + "type": "comment", + "insight": "The references show a strong focus on neural network foundations and deep learning research, including classic papers on GANs, transformers, and variational autoencoders.", + "content": "[11] A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit,\nL. Jones, A. N. Gomez, L. u. Kaiser, and I. Polosukhin,\ninAdvances in Neural Information Processing Systems,\nVol. 30, edited by I. Guyon, U. V. Luxburg, S. Bengio,\nH. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett\n(Curran Associates, Inc., 2017).\n[14] H. Xiao, K. Rasul, and R. Vollgraf, arXiv [cs.LG] (2017).\n[16] D.P.KingmaandM.Welling,Auto-Encoding Variational\nBayes(2022).\n[17] I. J. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu,\nD. Warde-Farley, S. Ozair, A. Courville, and Y. Ben-\nbio, inAdvances in Neural Information Processing Sys-\ntems, Vol. 27, edited by Z. Ghahramani, M. Welling,\nC. Cortes, N. Lawrence, and K. Weinberger (Curran As-\nsociates, Inc., 2014).\n[15] M. Heusel, H. Ramsauer, T. Unterthiner, B. Nessler,\nand S. Hochreiter, inAdvances in Neural Information Pro-\ncessing Systems, Vol. 30, edited by I. Guyon, U. V.\nLuxburg, S. Bengio, H. Wallach, R. Fergus, S. Vish-\nwanathan, and R. Garnett (Curran Associates, Inc.,\n2017).", + "attributes": [ + { + "attribute": "page_number", + "value": "10" + }, + { + "attribute": "topic_area", + "value": "Neural Network Foundations" + }, + { + "attribute": "key_papers", + "value": "Transformers, GANs, VAEs, Deep Learning" + } + ] + }, + { + "type": "fact", + "insight": "The bibliography includes significant quantum computing and neuromorphic research, with multiple references to applications in AI training and information processing.", + "content": "[28] C. Pratt, K. Ray, and J. Crutchfield,Dynamical Com-\nputing on the Nanoscale: Superconducting Circuits for\nThermodynamically-Efficient Classical Information Pro-\ncessing(2023).\n[29] G. Wimsatt, O.-P. Saira, A. B. Boyd, M. H. Matheny,\nS. Han, M. L. Roukes, and J. P. Crutchfield, Phys. Rev.\nRes.3, 033115 (2021).\n[30] S. H. Adachi and M. P. Henderson,Application of Quan-\ntum Annealing to Training of Deep Neural Networks\n(2015).\n[31] B. Sutton, K. Y. Camsari, B. Behin-Aein, and S. Datta,\nSci. Rep.7, 44370 (2017).\n[32] R. Faria, K. Y. Camsari, and S. Datta, IEEE Magn. Lett.\n8, 1 (2017).\n[33] S.Niazi, S.Cowdhury, N.A.Aadit, M.Mohseni, Y.Qin,\nand K. Y. Camsari, Nat. Electron.7, 610 (2024).\n[34] W. A. Borders, A. Z. Pervaiz, S. Fukami, K. Y. Camsari,\nH. Ohno, and S. Datta, Nature573, 390 (2019).", + "attributes": [ + { + "attribute": "page_number", + "value": "10" + }, + { + "attribute": "research_area", + "value": "Quantum Computing & Neuromorphic Computing" + }, + { + "attribute": "applications", + "value": "Neural Network Training, Information Processing" + } + ] + }, + { + "type": "fact", + "insight": "Denoising diffusion models learn to time-reverse a random process that converts data into simple noise", + "content": "Denoising diffusion models try to learn to time-reverse a random process that converts data into simple noise. Here, we will review some details on how these models work to support the analysis in the main text.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "11" + }, + { + "attribute": "topic", + "value": "Introduction" + } + ] + }, + { + "type": "fact", + "insight": "The forward process converts data distribution into noise through stochastic differential equations (continuous case) or Markov jump processes (discrete case)", + "content": "The forward process is a random process that is used to convert the data distribution into noise. This conversion into noise is achieved through a stochastic differential equation in the continuous-variable case and a Markov jump process in the discrete case.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "11" + }, + { + "attribute": "topic", + "value": "Forward Processes" + } + ] + }, + { + "type": "fact", + "insight": "In the continuous case, the typical forward process is Itô diffusion with the specific equation dX(t) = −X(t)dt + √2σdW", + "content": "In the continuous case, the typical choice of forward process is the Itô diffusion, dX(t) =−X(t)dt+ √2σdW where X(t) is a length N vector representing the state variable at time t, σ is a constant, and dW is a length N vector of independent Wiener processes.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "11" + }, + { + "attribute": "topic", + "value": "Continuous Variables" + }, + { + "attribute": "equation", + "value": "A1" + } + ] + }, + { + "type": "fact", + "insight": "The transition kernel defines how probability distribution evolves over time, given by Qt|0(x′|x) = P(X(t) =x ′|X(0) =x)", + "content": "The transition kernel for a random process defines how the probability distribution evolves in time, Qt|0(x′|x) =P(X(t) =x ′|X(0) =x)", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "11" + }, + { + "attribute": "topic", + "value": "Continuous Variables" + }, + { + "attribute": "equation", + "value": "A2" + } + ] + }, + { + "type": "fact", + "insight": "For the Itô diffusion case, the transition kernel follows a Gaussian distribution with mean µ=e^(-tx) and covariance matrix Σ=σ²I(1−e^(-2t))", + "content": "For the case of Eq. (A1) the transition kernel is, Qt+s|s(x′|x)∝e^−1/2 (x′−µ)^T Σ^−1(x′−µ) µ=e^−tx Σ =σ²I(1−e^−2t) this solution can be verified by direct substitution into the corresponding Fokker-Planck equation.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "11" + }, + { + "attribute": "topic", + "value": "Continuous Variables" + }, + { + "attribute": "equation", + "value": "A3-A5" + } + ] + }, + { + "type": "fact", + "insight": "The stationary distribution of the continuous case is zero-mean Gaussian noise with standard deviation σ", + "content": "In the limit of infinite time,µ→0andΣ→σ²I. Therefore, the stationary distribution of this process is zero-mean Gaussian noise with a standard deviation ofσ.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "11" + }, + { + "attribute": "topic", + "value": "Continuous Variables" + } + ] + }, + { + "type": "fact", + "insight": "Discrete variable dynamics are described by Markov jump processes with generator L, where dQt/dt = LQt", + "content": "The stochastic dynamics of some discrete variableXmay be described by the Markov jump process, dQt dt =LQ t where L is the generator of the dynamics, which is anM×Mmatrix that stores the transition rates between the various states.Q t is a lengthMvector that assigns a probability to each possible stateXmay take at timet.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "11" + }, + { + "attribute": "topic", + "value": "Discrete Variables" + }, + { + "attribute": "equation", + "value": "A6" + } + ] + }, + { + "type": "fact", + "insight": "The transition rates between discrete states are given by L[j, i] = γ(−(M−1)δ j,i + (1−δ j,i)), where δ is the Kronecker delta function", + "content": "The transition rate from theith state to thejth state is given by the matrix elementL[j, i], which here takes the particular form, L[j, i] =γ(−(M−1)δ j,i + (1−δ j,i)) whereδis used to indicate the Kronecker delta function. Eq. (A7) describes a random process where the probability per unit time to jump between any two states isγ.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "11" + }, + { + "attribute": "topic", + "value": "Discrete Variables" + }, + { + "attribute": "equation", + "value": "A7" + } + ] + }, + { + "type": "fact", + "insight": "The dynamics of Qt can be understood through eigenvalues and eigenvectors of L, given by Lvk = λkvk", + "content": "Since Eq. (A6) is linear, the dynamics ofQt can be understood entirely via the eigenvalues and eigenvectors ofL, Lvk =λ kvk", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "11" + }, + { + "attribute": "topic", + "value": "Discrete Variables" + }, + { + "attribute": "equation", + "value": "A8" + } + ] + }, + { + "type": "fact", + "insight": "Denoising diffusion models learn to time-reverse a random process that converts data into simple noise through stochastic differential equations or Markov jump processes", + "content": "Denoising diffusion models try to learn to time-reverse a random process that converts data into simple noise. Here, we will review some details on how these models work to support the analysis in the main text.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "12" + } + ] + }, + { + "type": "fact", + "insight": "Forward processes in diffusion models convert data distribution into noise using different mathematical approaches for continuous and discrete variables", + "content": "The forward process is a random process that is used to convert the data distribution into noise. This conversion into noise is achieved through a stochastic differential equation in the continuous-variable case and a Markov jump process in the discrete case.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "12" + } + ] + }, + { + "type": "fact", + "insight": "Continuous variable diffusion models use Itô diffusion equation dX(t) = -X(t)dt + √2σdW to transform data into noise", + "content": "In the continuous case, the typical choice of forward process is the Itô diffusion, dX(t) =−X(t)dt+√2σdW where X(t) is a length N vector representing the state variable at time t, σ is a constant, and dW is a length N vector of independent Wiener processes.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "12" + } + ] + }, + { + "type": "fact", + "insight": "The transition kernel Qt|0(x'|x) defines how probability distributions evolve over time in continuous diffusion models", + "content": "The transition kernel for a random process defines how the probability distribution evolves in time, Qt|0(x′|x) =P(X(t) =x ′|X(0) =x)", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "12" + } + ] + }, + { + "type": "fact", + "insight": "For Itô diffusion, the transition kernel follows a Gaussian distribution with parameters μ = e^(-t)x and Σ = σ^2 I (1−e^(-2t))", + "content": "For the case of Eq. (A1) the transition kernel is, Qt+s|s(x′|x)∝e −1/2 (x′−μ)T Σ−1(x′−μ) μ=e −tx Σ =σ2I 1−e −2t this solution can be verified by direct substitution into the corresponding Fokker-Planck equation.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "12" + } + ] + }, + { + "type": "fact", + "insight": "The stationary distribution of continuous diffusion processes is zero-mean Gaussian noise with standard deviation σ in the limit of infinite time", + "content": "In the limit of infinite time,μ→0andΣ→σ2I. Therefore, the stationary distribution of this process is zero-mean Gaussian noise with a standard deviation ofσ.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "12" + } + ] + }, + { + "type": "fact", + "insight": "Discrete variable diffusion models use Markov jump processes described by dQt/dt = LQt where L is the dynamics generator matrix", + "content": "The stochastic dynamics of some discrete variableXmay be described by the Markov jump process, dQt/dt =LQ t whereLis the generator of the dynamics, which is anM×Mmatrix that stores the transition rates between the various states.Q t is a lengthMvector that assigns a probability to each possible stateXmay take at timet.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "12" + } + ] + }, + { + "type": "fact", + "insight": "Discrete diffusion transition rates between states are given by L[j,i] = γ(-(M-1)δj,i + (1-δj,i)) where γ is the jump probability rate", + "content": "The transition rate from theith state to thejth state is given by the matrix elementL[j, i], which here takes the particular form, L[j, i] =γ(−(M−1)δj,i + (1−δj,i)) whereδis used to indicate the Kronecker delta function. Eq. (A7) describes a random process where the probability per unit time to jump between any two states isγ.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "12" + } + ] + }, + { + "type": "fact", + "insight": "Markov jump process dynamics can be analyzed through eigenvalues and eigenvectors of the generator matrix L", + "content": "Since Eq. (A6) is linear, the dynamics ofQt can be understood entirely via the eigenvalues and eigenvectors ofL, Lvk =λ kvk", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "12" + } + ] + }, + { + "type": "fact", + "insight": "Discrete diffusion processes have one stationary eigenvector with eigenvalue 0 and M-1 decaying modes with negative eigenvalues λj = -γM", + "content": "One eigenvector-eigenvalue pair(v0, λ0 = 0)corresponds to the unique stationary state ofL, with all entries ofv0 being equal to some constant (if normalized, thenv0[j] = 1/M for allj). The remaining eigenvectors are decaying modes associated with negative eigenvalues. These additionalM−1 eigenvectors take the form, vj[i] =−δi,0 +δi,j λj =−γM where Eq. (A9) and Eq. (A10) are valid forj∈[1, M−1]. Therefore, all solutions to this MJP decay exponentially to the uniform distribution with rateγM.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "Time evolution of discrete diffusion processes is given by Qt = e^(Lt)Q0 where the matrix exponential is evaluated through diagonalization", + "content": "The time-evolution ofQis given by the matrix exponential, Qt =e LtQ0.This matrix exponential is evaluated by diagonalizingL, eLt =P eDtP−1 where the columns ofPare theMeigenvectorsv k andDis a diagonal matrix of the eigenvaluesλk.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "Matrix elements of discrete diffusion time evolution follow specific exponential forms involving delta functions and exponential decay terms", + "content": "Using the solution for the eigenvalues and eigenvectors found above, we can solve for the matrix elements ofeLt, eLt [j, i] =δi,j (1 + (M−1)e −γMt/M) + (1−δi,j)(1−e −γMt/M) Using this solution, we can deduce an exponential form for the matrix elements ofeLt, eLt [j, i] = 1/Z(t) eΓ(t)δi,j Γ(t) = ln((1 + (M−1)e −γt)/(1−e −γt)) Z(t) = M/(1−e −γt)", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "For multiple independent discrete variables, the joint distribution dynamics uses Kronecker products of individual diffusion operators", + "content": "Now consider a process in which each element of the vector ofNdiscrete variablesXundergoes the dynamics described by Eq. (A6) independently. In that case, the differential equation describing the dynamics of the joint distributionQ t is, dQt/dt = NX k=1 (I1 ⊗ ··· ⊗ Lk ⊗. . . IN )Q t whereI j indicates the identity operator andLj the operator from Eq. (A7) acting on the subspace of thejth discrete variable.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "Time evolution of joint distributions for multiple discrete variables uses the Kronecker product of individual matrix exponentials", + "content": "The Kronecker product of the matrix exponentials gives the time-evolution of the joint distribution, eLt = NO k=1 eLkt", + "attributes": [ + { + "attribute": "section", + "value": "Appendix A: Denoising Diffusion Models" + }, + { + "attribute": "page", + "value": "13" + } + ] + }, + { + "type": "fact", + "insight": "For continuous diffusion models, the generator L corresponds to the Itô diffusion process and follows the Fokker-Planck equation structure with drift and diffusion terms.", + "content": "In the case that the forward process is an Itô diffusion,Lis the generator for the corresponding Fokker-Planck equation,\nL=−\nX\ni\n∂\n∂xi\nfi(x, t) +1\n2\nX\ni,j\n∂\n∂xi\n∂\n∂xj\nDij(t)(A29)", + "attributes": [ + { + "attribute": "equation", + "value": "(A29)" + }, + { + "attribute": "type", + "value": "continuous_diffusion" + } + ] + }, + { + "type": "fact", + "insight": "The adjoint operator L† for continuous diffusion models has a specific form with drift and diffusion terms, derived using integration by parts.", + "content": "Using Eq. (A28) and integration by parts, it can be shown that the adjoint operator is,\nL† =\nX\ni\nfi\n∂\n∂xi\n+ 1\n2\nX\ni,j\nDij\n∂\n∂xi\n∂\n∂xi\n(A30)", + "attributes": [ + { + "attribute": "equation", + "value": "(A30)" + }, + { + "attribute": "type", + "value": "adjoint_operator" + } + ] + }, + { + "type": "fact", + "insight": "The reverse process operator Lrev for continuous diffusion models can be reduced to a form involving a drift vector g and diffusion matrix D.", + "content": "By directly substituting Eq. (A30) into Eq. (A27) and simplifying,Lrev can be reduced to,\nLrev =\nX\ni\n∂\n∂xi\ngi + 1\n2\nX\ni,j\n∂\n∂xi\n∂\n∂xj\nDij (A31)", + "attributes": [ + { + "attribute": "equation", + "value": "(A31)" + }, + { + "attribute": "type", + "value": "reverse_operator" + } + ] + }, + { + "type": "fact", + "insight": "The drift vector gi in continuous diffusion models is defined as fi(x,t) minus a term involving the gradient of the diffusion matrix weighted by the probability distribution Qt(x).", + "content": "with the drift vectorg,\ngi(x, t) =fi(x, t)− 1\nQt(x)\nX\nj\n∂\n∂xj\n[Dij(x, t)Qt(x)](A32)", + "attributes": [ + { + "attribute": "equation", + "value": "(A32)" + }, + { + "attribute": "type", + "value": "drift_vector" + } + ] + }, + { + "type": "fact", + "insight": "For sufficiently small time steps Δt, the transition kernel in continuous diffusion models becomes Gaussian with mean μ and covariance matrix Σ defined by the drift and diffusion terms.", + "content": "If∆tis chosen to be sufficiently small, Eq. (A32) can be linearized and the transition kernel is Gaussian,\nQt|t+∆t(x′|x)∝exp\n(\n−1\n2(x−µ) T Σ−1(x−µ)\n\n)\n(A33)", + "attributes": [ + { + "attribute": "equation", + "value": "(A33)" + }, + { + "attribute": "type", + "value": "gaussian_kernel" + } + ] + }, + { + "type": "fact", + "insight": "The Gaussian mean vector μ is defined as x + Δt gi(x,t) and the covariance matrix Σ is Δt D(t) for continuous diffusion models with small Δt.", + "content": "µ=x+ ∆t g i(x, t)(A34)\nΣ = ∆t D(t)(A35)", + "attributes": [ + { + "attribute": "equation", + "value": "(A34)-(A35)" + }, + { + "attribute": "type", + "value": "gaussian_parameters" + } + ] + }, + { + "type": "comment", + "insight": "Continuous diffusion models can achieve arbitrary approximation power in the small Δt limit by using neural networks to define the mean vector in Gaussian transitions.", + "content": "Therefore, one can build a continuous diffusion model with arbitrary approximation power by working in the small∆t\nlimit and approximating the reverse process using a Gaussian distribution with a neural network defining the mean\nvector [1, 2].", + "attributes": [ + { + "attribute": "method", + "value": "neural_network_approximation" + }, + { + "attribute": "scope", + "value": "continuous_models" + } + ] + }, + { + "type": "fact", + "insight": "For discrete diffusion models, the operator L has a tensor product form that guarantees L(x′, x) = 0 for any vectors with Hamming distance greater than one.", + "content": "In a discrete diffusion model,Lis given by Eq. (A17). This tensor product form forLguarantees thatL(x′, x) = 0\nfor any vectorsx ′ andxthat have a Hamming distance greater than one (which means they have at leastN−1\nmatching elements).", + "attributes": [ + { + "attribute": "equation", + "value": "(A17)" + }, + { + "attribute": "type", + "value": "discrete_operator" + } + ] + }, + { + "type": "comment", + "insight": "Neural networks can be effectively used in discrete diffusion models to approximate ratios of data distributions for neighboring states, enabling arbitrarily good approximations to the reverse process.", + "content": "As such, in discrete diffusion models, neural networks trained to approximate ratios of the data distribution QT−s (x′)\nQT−s (x) for neighboringx ′ andxcan be used to implement an arbitrarily good approximation to the\nactual reverse process [3].", + "attributes": [ + { + "attribute": "method", + "value": "neural_network_approximation" + }, + { + "attribute": "scope", + "value": "discrete_models" + }, + { + "attribute": "reference", + "value": "[3]" + } + ] + }, + { + "type": "fact", + "insight": "The adjoint operator L† for continuous diffusion models with Itô diffusion processes is mathematically defined as L† = Σᵢ fᵢ ∂/∂xᵢ + ½ Σᵢⱼ Dᵢⱼ ∂²/∂xᵢ∂xⱼ, where D is a symmetric matrix not dependent on x.", + "content": "Using Eq. (A28) and integration by parts, it can be shown that the adjoint operator is,\nL† = Σᵢ fᵢ ∂/∂xᵢ + ½ Σᵢⱼ Dᵢⱼ ∂/∂xᵢ ∂/∂xᵢ (A30)", + "attributes": [ + { + "attribute": "section", + "value": "Continuous variables" + }, + { + "attribute": "equation", + "value": "(A30)" + }, + { + "attribute": "page", + "value": "15" + } + ] + }, + { + "type": "fact", + "insight": "In the small Δt limit, continuous diffusion models can be constructed using Gaussian distributions with neural networks defining the mean vector, providing arbitrary approximation power.", + "content": "Therefore, one can build a continuous diffusion model with arbitrary approximation power by working in the small Δt limit and approximating the reverse process using a Gaussian distribution with a neural network defining the mean vector [1, 2].", + "attributes": [ + { + "attribute": "section", + "value": "Continuous variables" + }, + { + "attribute": "reference", + "value": "[1, 2]" + }, + { + "attribute": "page", + "value": "15" + } + ] + }, + { + "type": "fact", + "insight": "The diffusion loss LDN(θ) is formulated as minimizing the distributional distance between the joint distributions of forward process Q₀,...,T and learned reverse process approximation Pθ₀,...,T, which can be simplified to a layerwise form for Markovian processes.", + "content": "LDN (θ) = D(Q₀,...,T (·)||Pθ₀,...,T (·)) (A36)\nthe Markovian nature of Q can be taken advantage of to simplify Eq. (A36) into a layerwise form,\nLDN (θ) +C=− Σᵗ₌₁ᵀ E_Q(xₜ₋₁,xₜ) [log (Pθ(xₜ₋₁|xₜ)](A37)", + "attributes": [ + { + "attribute": "equation", + "value": "(A36), (A37)" + }, + { + "attribute": "section", + "value": "The Diffusion Loss" + }, + { + "attribute": "page", + "value": "16" + } + ] + }, + { + "type": "fact", + "insight": "For discrete diffusion models, neural networks trained to approximate ratios of data distributions QT₋ₛ(x')/QT₋ₛ(x) for neighboring states x' and x can implement arbitrarily good approximations to the actual reverse process.", + "content": "As such, in discrete diffusion models, neural networks trained to approximate ratios of the data distribution QT−s (x′)\nQT−s (x) for neighboringx ′ andxcan be used to implement an arbitrarily good approximation to the actual reverse process [3].", + "attributes": [ + { + "attribute": "reference", + "value": "[3]" + }, + { + "attribute": "section", + "value": "Discrete variables" + }, + { + "attribute": "page", + "value": "15" + } + ] + }, + { + "type": "fact", + "insight": "Conditional generation tasks like generating MNIST digits with specific class labels can be implemented by concatenating the target data with one-hot encoded labels and treating them as augmented training data for the denoising model.", + "content": "In principle, this is very simple: we concatenate the target (in our case, the images) and a one-hot encoding of the labels into a contiguous binary vector and treat that whole thing as our training data on which we train the denoising model as described above.", + "attributes": [ + { + "attribute": "application", + "value": "MNIST digit generation" + }, + { + "attribute": "section", + "value": "Conditional Generation" + }, + { + "attribute": "page", + "value": "17" + } + ] + }, + { + "type": "opinion", + "insight": "The energy landscape of EBM-based approximations to reverse processes becomes simpler as the forward process timestep decreases, making sampling easier despite potential training challenges with noised labels during conditional inference.", + "content": "However, during conditional inference, the models will have their label nodes clamped to an unnoised labell0, and they may not know how this should influence the generated image (and this problem would only be exacerbated if we clamped to a noised label instead).\nThis issue can be mitigated by using a rateγX when noising image entries in the training data and a different rateγL for noising label entries.", + "attributes": [ + { + "attribute": "recommendation", + "value": "Use different noise rates for images and labels" + }, + { + "attribute": "section", + "value": "Conditional Generation" + }, + { + "attribute": "page", + "value": "17" + } + ] + }, + { + "type": "comment", + "insight": "Figure 7 demonstrates that as λ increases (representing smaller forward process timesteps), the energy landscape transitions from a strongly bimodal distribution to a simple Gaussian centered at xt, making the distribution much easier to sample from.", + "content": "The energy landscape is bimodal atλ= 0and gradually becomes distorted towards an unimodal distribution centered atx t asλincreases. This reshaping is intuitive, as shortening the forward process timestep should more strongly constrainx t−1 tox t.", + "attributes": [ + { + "attribute": "visualization", + "value": "FIG. 7" + }, + { + "attribute": "section", + "value": "Simplification of the Energy Landscape" + }, + { + "attribute": "page", + "value": "16" + } + ] + }, + { + "type": "fact", + "insight": "For energy-based models (EBMs) where Pθ(xₜ₋₁|xₜ) has no closed-form expression, Monte Carlo estimators must be employed to approximate the gradients of the diffusion loss, with the gradient derived as -Σᵗ₌₁ᵀ E_Q(xₜ₋₁,xₜ)[∇θ log Pθ(xₜ₋₁|xₜ)].", + "content": "∇θLDN (θ) =−\nTX\nt=1\nEQ(xt−1,xt)\n[∇θ log\nPθ(xt−1|xt)\n]\n(A38)", + "attributes": [ + { + "attribute": "equation", + "value": "(A38)" + }, + { + "attribute": "section", + "value": "Monte-Carlo gradient estimator" + }, + { + "attribute": "page", + "value": "16" + } + ] + }, + { + "type": "fact", + "insight": "The diffusion loss function LDN(θ) is defined as the distributional distance between joint distributions of forward process Q₀,...,T and reverse process Pθ₀,...,T (Equation A36)", + "content": "LDN (θ) =D\nQ0,...,T (·)||Pθ\n0,...,T (·)\n(A36)", + "attributes": [ + { + "attribute": "equation", + "value": "A36" + }, + { + "attribute": "section", + "value": "3" + } + ] + }, + { + "type": "fact", + "insight": "The Markovian nature of Q allows simplification of the diffusion loss into a layerwise form (Equation A37)", + "content": "LDN (θ) +C=−\nTX\nt=1\nEQ(xt−1,xt) [log (Pθ(xt−1|xt)](A37)", + "attributes": [ + { + "attribute": "equation", + "value": "A37" + }, + { + "attribute": "section", + "value": "3" + } + ] + }, + { + "type": "fact", + "insight": "For denoising algorithms in the infinitesimal limit, gradients of LDN can be computed exactly due to the simple form of Pθ", + "content": "For denoising algorithms that operate in the infinitesimal limit, the simple form of Pθ allows forLDN and its gradients to be computed exactly.", + "attributes": [ + { + "attribute": "section", + "value": "3" + } + ] + }, + { + "type": "fact", + "insight": "When Pθ(xt-1|xt) is an EBM (Energy-Based Model), no closed-form expression exists for ∇θLDN(θ), requiring Monte Carlo estimation", + "content": "In the case wherePθ\nxt−1|xt\n\nis an EBM, there exists no simple closed-form expression for∇θLDN (θ). In that case, one must employ a Monte Carlo estimator to approximate the gradient.", + "attributes": [ + { + "attribute": "section", + "value": "3a" + }, + { + "attribute": "model_type", + "value": "EBM" + } + ] + }, + { + "type": "fact", + "insight": "The Monte Carlo gradient estimator for diffusion loss is derived as shown in Equation (A38)", + "content": "∇θLDN (θ) =−\nTX\nt=1\nEQ(xt−1,xt)\n ∇θ log\n Pθ(xt−1|xt)\n \n (A38)", + "attributes": [ + { + "attribute": "equation", + "value": "A38" + }, + { + "attribute": "section", + "value": "3a" + } + ] + }, + { + "type": "fact", + "insight": "With EBM parameterization, the gradient of log-likelihood can be simplified using latent variables as shown in Equation (A39)", + "content": "∇θ log\n Pθ(xt−1|xt)\n \n =E Pθ(xt−1,zt−1|xt)\n ∇θEm\n t−1\n \n −E Pθ(zt−1|xt−1,xt)\n ∇θEm\n t−1\n \n (A39)", + "attributes": [ + { + "attribute": "equation", + "value": "A39" + }, + { + "attribute": "section", + "value": "3a" + }, + { + "attribute": "model_type", + "value": "EBM" + } + ] + }, + { + "type": "fact", + "insight": "Smaller forward process timesteps lead to simpler energy landscapes in EBM-based approximations", + "content": "As the forward process timestep is made smaller, the energy landscape of the EBM-based approximation to the reverse process becomes simpler.", + "attributes": [ + { + "attribute": "section", + "value": "4" + }, + { + "attribute": "model_type", + "value": "EBM" + } + ] + }, + { + "type": "fact", + "insight": "The marginal energy function example shows Eθ_t-1(xt-1) = (xt-1² - 1)² (Equation A40)", + "content": "Eθ\nt−1 (xt−1) =\n x2\n t−1 −1\n \n 2\n(A40)", + "attributes": [ + { + "attribute": "equation", + "value": "A40" + }, + { + "attribute": "section", + "value": "4" + } + ] + }, + { + "type": "fact", + "insight": "The forward process energy function for Gaussian diffusion is Ef_t-1(xt-1, xt) = λ((xt-1/xt) - 1)² (Equation A41)", + "content": "Ef\nt−1 (xt−1, xt) =λ\n xt−1\n xt\n −1\n \n 2\n(A41)", + "attributes": [ + { + "attribute": "equation", + "value": "A41" + }, + { + "attribute": "section", + "value": "4" + }, + { + "attribute": "process_type", + "value": "Gaussian diffusion" + } + ] + }, + { + "type": "fact", + "insight": "Parameter λ scales inversely with forward process timestep size, approaching infinity as Δt → 0", + "content": "The parameterλscales inversely with the size of the forward process timestep; that is,lim\n∆t→0\nλ=∞.", + "attributes": [ + { + "attribute": "section", + "value": "4" + }, + { + "attribute": "parameter", + "value": "λ" + } + ] + }, + { + "type": "fact", + "insight": "The reverse process conditional energy landscape is Eθ_t-1 + Ef_t-1", + "content": "The reverse process conditional energy landscape is thenEθ\nt−1 +E f\nt−1.", + "attributes": [ + { + "attribute": "section", + "value": "4" + } + ] + }, + { + "type": "opinion", + "insight": "The energy landscape transformation from bimodal at λ=0 to unimodal centered at xt as λ increases is intuitive because shorter forward process timesteps more strongly constrain xt-1 to xt", + "content": "This reshaping is intuitive, as shortening the forward process timestep should more strongly constrainx t−1 tox t.", + "attributes": [ + { + "attribute": "section", + "value": "4" + }, + { + "attribute": "sentiment", + "value": "intuitive" + } + ] + }, + { + "type": "comment", + "insight": "Figure 7 illustrates the effect of λ on the energy landscape, showing the gradual distortion from bimodal to unimodal distribution", + "content": "The effect ofλon this is shown in Fig. 7.\nThe energy landscape is bimodal atλ= 0and gradually becomes distorted towards an unimodal distribution centered atx t asλincreases.", + "attributes": [ + { + "attribute": "section", + "value": "4" + }, + { + "attribute": "reference", + "value": "Figure 7" + }, + { + "attribute": "observation", + "value": "bimodal to unimodal transition" + } + ] + }, + { + "type": "fact", + "insight": "Experimental parameter ranges for good conditional generation performance: γL ∈ [0.1,0.3] and γX ∈ [0.7,1.5] for models with 4-12 steps", + "content": "Experimentally, we observed that settings in the rangesγL ∈[0.1,0.3]andγ X ∈[0.7,1.5](for models with four to 12 steps) yielded good conditional generation performance while avoiding the freezing problem.", + "attributes": [ + { + "attribute": "methodology", + "value": "experimental" + }, + { + "attribute": "parameter_range", + "value": "γL: [0.1,0.3], γX: [0.7,1.5]" + }, + { + "attribute": "model_complexity", + "value": "4-12 steps" + } + ] + }, + { + "type": "fact", + "insight": "Theoretical equivalence between learned energy functions and true marginal distribution in DTM", + "content": "If a DTM is trained to match the conditional distribution of the reverse process perfectly, the learned energy functionE θ t−1 is the energy function of the true marginal distribution, that is,Eθ t−1(x)∝logQ(x t−1).", + "attributes": [ + { + "attribute": "theoretical_basis", + "value": "Bayes' rule" + }, + { + "attribute": "model_type", + "value": "DTM (Diffusion Transition Model)" + }, + { + "attribute": "mathematical_relationship", + "value": "Eθ t−1(x)∝logQ(x t−1)" + } + ] + }, + { + "type": "fact", + "insight": "Hardware architecture approach for EBMs using Probabilistic Graphical Models", + "content": "In this work, we focus on a hardware architecture for EBMs that are naturally expressed as Probabilistic Graphical Models (PGMs). In a PGM-EBM, the random variables involved in the model map to the nodes of a graph, which are connected by edges that indicate dependence between variables.", + "attributes": [ + { + "attribute": "architecture_type", + "value": "PGM-based EBM hardware" + }, + { + "attribute": "model_representation", + "value": "Probabilistic Graphical Models" + }, + { + "attribute": "computational_approach", + "value": "modular sampling" + } + ] + }, + { + "type": "fact", + "insight": "Energy efficiency benefits of local PGM samplers over Von Neumann architectures", + "content": "Since the sampling circuits only communicate locally, this type of computer will spend significantly less energy on communication than one built on a Von-Neumann-like architecture, which constantly shuttles data between compute and memory.", + "attributes": [ + { + "attribute": "efficiency_advantage", + "value": "reduced communication energy" + }, + { + "attribute": "architecture_comparison", + "value": "local vs Von Neumann" + }, + { + "attribute": "computational_model", + "value": "compute-in-memory" + } + ] + }, + { + "type": "fact", + "insight": "Gibbs sampling algorithm definition for PGM joint distribution sampling", + "content": "Formally, the algorithm that defines this modular sampling procedure for PGMs is called Gibbs sampling. In Gibbs sampling, samples are drawn from the joint distributionp(x1, x2, . . . , xN )by iteratively updating the state of each node conditioned on the current state of its neighbors. For theith node, this means sampling from the distribution, xi[t+ 1]∼p(x i|nb(xi)[t]).", + "attributes": [ + { + "attribute": "algorithm", + "value": "Gibbs sampling" + }, + { + "attribute": "sampling_method", + "value": "iterative node conditioning" + }, + { + "attribute": "mathematical_formulation", + "value": "xi[t+1]∼p(xi|nb(xi)[t])" + } + ] + }, + { + "type": "fact", + "insight": "Experimental settings with gamma parameters γL ∈[0.1,0.3] and γX ∈[0.7,1.5] for models with 4-12 steps achieved good conditional generation performance while avoiding the freezing problem.", + "content": "Experimentally, we observed that settings in the rangesγL ∈[0.1,0.3]andγ X ∈[0.7,1.5](for models with four to 12 steps) yielded good conditional generation performance while avoiding the freezing problem.", + "attributes": [ + { + "attribute": "source", + "value": "Experimental observation" + }, + { + "attribute": "model_range", + "value": "4-12 steps" + }, + { + "attribute": "gamma_L", + "value": "[0.1,0.3]" + }, + { + "attribute": "gamma_X", + "value": "[0.7,1.5]" + }, + { + "attribute": "performance", + "value": "good conditional generation" + }, + { + "attribute": "issue_avoided", + "value": "freezing problem" + } + ] + }, + { + "type": "fact", + "insight": "When a DTM is trained to perfectly match the conditional distribution of the reverse process, the learned energy function Eθ t−1 corresponds to the energy function of the true marginal distribution, satisfying Eθ t−1(x)∝logQ(x t−1).", + "content": "If a DTM is trained to match the conditional distribution of the reverse process perfectly, the learned energy functionE θ t−1 is the energy function of the true marginal distribution, that is,Eθ t−1(x)∝logQ(x t−1).", + "attributes": [ + { + "attribute": "theoretical_basis", + "value": "Bayes' rule application" + }, + { + "attribute": "model_type", + "value": "DTM (Diffusion Transition Model)" + }, + { + "attribute": "condition", + "value": "perfect match with true reverse process" + }, + { + "attribute": "mathematical_relationship", + "value": "Eθ t−1(x)∝logQ(x t−1)" + } + ] + }, + { + "type": "comment", + "insight": "Probabilistic Graphical Models (PGMs) provide a natural basis for hardware architecture for EBMs because they can be sampled using modular procedures that respect graph structure, enabling efficient hardware implementation with local communication.", + "content": "PGMs form a natural basis for a hardware architecture because they can be sampled using a modular procedure that respects the graph's structure. Specifically, the state of a PGM can be updated by iteratively stepping through each node of the graph and resampling one variable at a time, using only information about the current node and its immediate neighbors.", + "attributes": [ + { + "attribute": "model_type", + "value": "PGM-EBM (Probabilistic Graphical Model - Energy-Based Model)" + }, + { + "attribute": "sampling_method", + "value": "Modular procedure respecting graph structure" + }, + { + "attribute": "advantage", + "value": "Local communication, efficient hardware implementation" + } + ] + }, + { + "type": "comment", + "insight": "Compute-in-memory approaches using local PGM sampling circuits can significantly reduce energy consumption compared to Von Neumann architectures by minimizing data communication between compute and memory components.", + "content": "This localPGMsampler representsa type of compute-in-memory approach, where the stateof the sampling program is spatially distributed throughout the array of sampling circuitry. Since the sampling circuits only communicate locally, this type of computer will spend significantly less energy on communication than one built on a Von Neumann-like architecture, which constantly shuttles data between compute and memory.", + "attributes": [ + { + "attribute": "architecture_type", + "value": "Compute-in-memory" + }, + { + "attribute": "advantage", + "value": "Significantly less energy on communication" + }, + { + "attribute": "comparison", + "value": "vs Von Neumann architecture" + } + ] + }, + { + "type": "fact", + "insight": "Gibbs sampling is the formal algorithm that defines the modular sampling procedure for PGMs, where samples are drawn from the joint distribution by iteratively updating each node conditioned on its neighbors.", + "content": "Formally, the algorithm that defines this modular sampling procedure for PGMs is called Gibbs sampling. In Gibbs sampling, samples are drawn from the joint distributionp(x1, x2, . . . , xN )by iteratively updating the state of each node conditioned on the current state of its neighbors. For theith node, this means sampling from the distribution, xi[t+ 1]∼p(x i|nb(xi)[t]).", + "attributes": [ + { + "attribute": "algorithm_name", + "value": "Gibbs sampling" + }, + { + "attribute": "sampling_procedure", + "value": "Iterative node updates conditioned on neighbors" + }, + { + "attribute": "mathematical_form", + "value": "xi[t+ 1]∼p(x i|nb(xi)[t])" + } + ] + }, + { + "type": "fact", + "insight": "Chromatic Gibbs Sampling enables parallel updates of graph nodes by grouping them into color classes where nodes of the same color do not neighbor each other, allowing simultaneous state updates.", + "content": "Since each node's update distribution only depends on the state of its neighbors and because nodes of the same color do not neighbor each other, they can all be updated in parallel.", + "attributes": [ + { + "attribute": "section", + "value": "Chromatic Gibbs Sampling" + }, + { + "attribute": "page", + "value": "19" + } + ] + }, + { + "type": "fact", + "insight": "Hardware acceleration of Gibbs sampling requires conditional updates to be efficiently implementable in the target hardware substrate, limiting the types of joint distributions that can be sampled.", + "content": "The primary constraint around building a hardware device that implements Gibbs sampling is that the conditional update given in Eq. (B1) must be efficiently implementable. Generally, this means that one wants it to take a form that is 'natural' to the hardware substrate being used to build the computer.", + "attributes": [ + { + "attribute": "section", + "value": "Quadratic EBMs" + }, + { + "attribute": "page", + "value": "19" + } + ] + }, + { + "type": "fact", + "insight": "Quadratic EBMs have energy functions that are quadratic in model variables, leading to conditional updates that are efficient to implement in various hardware types through simple sampling circuits biased by linear functions.", + "content": "Quadratic EBMs have energy functions that are quadratic in the model's variables, which generally leads to conditional updates computed by biasing a simple sampling circuit (Bernoulli, categorical, Gaussian, etc.) with the output of a linear function of the neighbor states and the model parameters. These simple interactions are efficient to implement in various types of hardware.", + "attributes": [ + { + "attribute": "section", + "value": "Quadratic EBMs" + }, + { + "attribute": "page", + "value": "19" + } + ] + }, + { + "type": "fact", + "insight": "Potts models extend Boltzmann machines to k-state variables, requiring softmax sampling circuits for hardware implementation rather than simpler Bernoulli sampling.", + "content": "Therefore, to build a hardware device that samples from Potts models using Gibbs sampling, one would have to build a softmax sampling circuit parameterized by a linear function of the model weights and neighbor states. Potts model sampling is slightly more complicated than Boltzmann machine sampling, but it is likely possible.", + "attributes": [ + { + "attribute": "section", + "value": "Potts models" + }, + { + "attribute": "page", + "value": "20" + } + ] + }, + { + "type": "fact", + "insight": "Gaussian-Bernoulli EBMs are more challenging to implement in hardware than discrete models because they require handling continuous signals between nodes, which can be done through discrete embedding or analog signaling but with significant overhead.", + "content": "Hardware implementations of Gaussian-Bernoulli EBMs are more difficult than the strictly discrete models because the signals being passed during conditional sampling of the binary variables are continuous. To pass these continuous values, they must either be embedded into several discrete variables or an analog signaling system must be used. Both of these solutions would incur significant overhead compared to the purely discrete models.", + "attributes": [ + { + "attribute": "section", + "value": "Gaussian-Bernoulli EBMs" + }, + { + "attribute": "page", + "value": "20" + } + ] + }, + { + "type": "fact", + "insight": "The denoising architecture in this work uses separate implementations for forward process and marginal energy functions, with the forward process implemented using simple pairwise couplings and the marginal energy function implemented using a grid-based Boltzmann machine.", + "content": "The denoising models used in this work exclusively modeled distributions of binary variables. The reverse process energy function (Eq. 7 in the main text) was implemented using a Boltzmann machine. The forward process energy functionEf t−1 was implemented using a simple set of pairwise couplings betweenxt (blue nodes) andxt−1 (green nodes). The marginal energy functionEθ t−1 was implemented using a latent variable model (latent nodes are drawn in orange) with a sparse, local coupling structure.", + "attributes": [ + { + "attribute": "section", + "value": "A hardware architecture for denoising" + }, + { + "attribute": "page", + "value": "21" + } + ] + }, + { + "type": "fact", + "insight": "The marginal energy function implementation uses a grid graph with nearest-neighbor and long-range skip connections, where some nodes represent data variables and others represent latent variables, creating a deep Boltzmann machine with sparse connectivity.", + "content": "Within the grid, we randomly choose some subset of the nodes to represent the data variablesxt−1. The remaining nodes then implement the latent variablezt−1. The grid is, therefore, a deep Boltzmann machine with a sparse connectivity structure and multiple hidden layers.", + "attributes": [ + { + "attribute": "section", + "value": "Implementation of the marginal energy function" + }, + { + "attribute": "page", + "value": "21" + } + ] + }, + { + "type": "opinion", + "insight": "The authors believe that while more complex than discrete models, Potts model sampling is likely implementable in hardware, though they suggest the experiments focused on simpler Boltzmann machines.", + "content": "Potts model sampling is slightly more complicated than Boltzmann machine sampling, but it is likely possible.", + "attributes": [ + { + "attribute": "section", + "value": "Potts models" + }, + { + "attribute": "page", + "value": "20" + } + ] + }, + { + "type": "comment", + "insight": "The document provides a comprehensive overview of different hardware-acceleratable EBM architectures, with detailed mathematical formulations for each type and specific implementation considerations.", + "content": "Here, we will touch on a few other types of quadratic EBM that are more general. Although the experiments in this paper focused on Boltzmann machines, they could be trivially extended to these more expressive classes of distributions.", + "attributes": [ + { + "attribute": "section", + "value": "Introduction to quadratic EBMs" + }, + { + "attribute": "page", + "value": "20" + } + ] + }, + { + "type": "fact", + "insight": "Potts models generalize Boltzmann machines to k-state variables using one-hot encoding where each variable xi has exactly one state xi^m = 1 and others are 0.", + "content": "xi^m is a one-hot encoding of the state of variable xi, xi^m ∈ {0,1} (B5) ∑^M_m=1 xi^m = 1 (B6) which implies that xi^m = 1 for a single value of m, and is zero otherwise.", + "attributes": [ + { + "attribute": "model_type", + "value": "Potts" + }, + { + "attribute": "encoding", + "value": "one-hot" + } + ] + }, + { + "type": "fact", + "insight": "Potts models have a softmax distribution for individual variables conditioned on their Markov blanket, which reduces to a simpler form when weight matrix J has symmetry J_ij^mn = J_ji^nm.", + "content": "p(xi^m = 1|mb(xi)) ∝ 1/Z e^{-θ_i^m} (B9) θ_i^m = β(2∑_{j∈mb(xi),n} J_ij^mn x_j^n + h_i^m) (B10)", + "attributes": [ + { + "attribute": "distribution_type", + "value": "softmax" + }, + { + "attribute": "condition", + "value": "symmetric weights" + } + ] + }, + { + "type": "comment", + "insight": "Hardware implementation of Potts models would require building a softmax sampling circuit parameterized by linear functions of model weights and neighbor states, which is more complex than Boltzmann machine sampling but likely feasible.", + "content": "Therefore, to build a hardware device that samples from Potts models using Gibbs sampling, one would have to build a softmax sampling circuit parameterized by a linear function of the model weights and neighbor states. Potts model sampling is slightly more complicated than Boltzmann machine sampling, but it is likely possible.", + "attributes": [ + { + "attribute": "implementation", + "value": "hardware" + }, + { + "attribute": "complexity", + "value": "moderate" + }, + { + "attribute": "feasibility", + "value": "likely possible" + } + ] + }, + { + "type": "fact", + "insight": "Gaussian-Bernoulli EBMs extend Boltzmann machines to continuous, binary mixtures and can handle continuous-continuous, binary-binary, and binary-continuous interactions.", + "content": "Gaussian-Bernoulli EBMs extend Boltzmann machines to continuous, binary mixtures. In general, this type of model can have continuous-continuous, binary-binary, and binary-continuous interactions.", + "attributes": [ + { + "attribute": "model_type", + "value": "Gaussian-Bernoulli" + }, + { + "attribute": "extensions", + "value": "continuous, binary mixtures" + }, + { + "attribute": "interaction_types", + "value": "three types" + } + ] + }, + { + "type": "fact", + "insight": "Quadratic EBMs beyond Boltzmann machines could trivially extend the experiments in this paper, though the focus was specifically on Boltzmann machines.", + "content": "Although the experiments in this paper focused on Boltzmann machines, they could be trivially extended to these more expressive classes of distributions.", + "attributes": [ + { + "attribute": "scope", + "value": "research experiments" + }, + { + "attribute": "extendibility", + "value": "trivial" + }, + { + "attribute": "focus", + "value": "Boltzmann machines" + } + ] + }, + { + "type": "fact", + "insight": "The hardware denoising architecture uses a grid subdivided into visible nodes (representing variables x_{t-1}) and latent nodes (representing z_{t-1}), with blue nodes carrying values from previous denoising steps that remain fixed during Gibbs sampling.", + "content": "A graph for hardware denoising. The grid is subdivided at random into visible (green) nodes, representing the variablesx t−1, and latent (orange) nodes, representingzt−1. Each visible nodext−1 j is coupled to a (blue) node carrying the value from the previous step of denoisingxt j (note that these blue nodes stay fixed throughout the Gibbs sampling).", + "attributes": [ + { + "attribute": "source", + "value": "Fig. 9b description" + }, + { + "attribute": "section", + "value": "Hardware Architecture" + } + ] + }, + { + "type": "fact", + "insight": "The variational approximation to the reverse process conditional uses an energy function that combines the forward process energy function and the marginal energy function, implemented by adding nodes to the grid that are connected pairwise to data nodes.", + "content": "As explicitly stated in Eq. 7 of the article, our variational approximation to the reverse process conditional has an energy function that is the sum of the forward process energy function and the marginal energy function. Physically, this corresponds to adding nodes to our grid that implementxt, which are connected pairwise to the data nodes implementingx t−1 via the coupling defined in Eq. (C1).", + "attributes": [ + { + "attribute": "source", + "value": "Section after Table I" + }, + { + "attribute": "equation", + "value": "Eq. 7" + } + ] + }, + { + "type": "fact", + "insight": "The hardware architecture uses a random number generator (RNG) circuit that produces random bits at approximately 10MHz using approximately 350aJ of energy per bit.", + "content": "We provide experimental measurements of our novel RNG circuitry in the main text, which establish that random bits can be produced at a rate ofτ−1 rng ≈10MHz using∼350aJ of energy per bit.", + "attributes": [ + { + "attribute": "source", + "value": "Appendix D: Energetic analysis" + }, + { + "attribute": "measurement", + "value": "Experimental" + }, + { + "attribute": "performance_metric", + "value": "10MHz, 350aJ/bit" + } + ] + }, + { + "type": "fact", + "insight": "The sampling cell design utilizes a linear analog circuit to combine neighboring states and model weights, producing a control voltage for an RNG that generates biased random bits based on a sigmoidal function of the control voltage.", + "content": "The design considered here utilizes a linear analog circuit to combine the neighboring states and model weights, producing a control voltage for an RNG. This RNG then produces a random bit that is biased by a sigmoidal function of the control voltage. This updated state is then broadcast back to the neighbors.", + "attributes": [ + { + "attribute": "source", + "value": "Appendix D: Sampling cell design" + }, + { + "attribute": "component", + "value": "Linear analog circuit + RNG" + } + ] + }, + { + "type": "comment", + "insight": "The document shows different connectivity patterns for various graph degrees (G8, G12, G16, G20, G24) with specific edge connections, indicating systematic scaling of the hardware architecture.", + "content": "Pattern Connectivity\nG8 (0,1),(4,1)\nG12 (0,1),(4,1),(9,10)\nG16 (0,1),(4,1),(8,7),(14,9)\nG20 (0,1),(4,1),(3,6),(8,7),(14,9)\nG24 (0,1),(1,2),(4,1),(3,6),(8,7),(14,9)", + "attributes": [ + { + "attribute": "source", + "value": "Table I" + }, + { + "attribute": "type", + "value": "Connectivity patterns" + } + ] + }, + { + "type": "comment", + "insight": "The RNG circuit output exhibits random wandering between high and low states, with the bias being a sigmoidal function of the control voltage, which is described as a critical feature for the system's operation.", + "content": "Fig. 15 (a) shows an output voltage waveform from the RNG circuit. It wanders randomly between high and low states. Critically, the bias of the RNG circuit (the probability of finding it in the high or low state) is a sigmoidal function of its control voltage, which allows", + "attributes": [ + { + "attribute": "source", + "value": "End of page 21" + }, + { + "attribute": "reference", + "value": "Fig. 15(a)" + } + ] + }, + { + "type": "fact", + "insight": "Hardware denoising architecture uses a grid-based connectivity pattern that is repeated for every cell in the grid.", + "content": "Our hardware denoising architecture (a)An example of a possible connectivity pattern as specified in Table. I. For clarity, the pattern is illustrated as applied to a single cell; however, in reality, the pattern is repeated for every cell in the grid.", + "attributes": [ + { + "attribute": "section", + "value": "Hardware Architecture Overview" + }, + { + "attribute": "reference", + "value": "Fig. 9(a)" + } + ] + }, + { + "type": "fact", + "insight": "The architecture implements visible nodes representing x_{t-1} and latent nodes representing z_{t-1}, with coupling between data nodes and previous denoising states.", + "content": "(b)A graph for hardware denoising. The grid is subdivided at random into visible (green) nodes, representing the variables x_{t-1}, and latent (orange) nodes, representing z_{t-1}. Each visible node x_{t-1}_j is coupled to a (blue) node carrying the value from the previous step of denoising x^j_t (note that these blue nodes stay fixed throughout the Gibbs sampling).", + "attributes": [ + { + "attribute": "section", + "value": "Hardware Architecture Overview" + }, + { + "attribute": "reference", + "value": "Fig. 9(b)" + } + ] + }, + { + "type": "fact", + "insight": "The variational approximation energy function is the sum of forward process energy function and marginal energy function.", + "content": "As explicitly stated in Eq. 7 of the article, our variational approximation to the reverse process conditional has an energy function that is the sum of the forward process energy function and the marginal energy function.", + "attributes": [ + { + "attribute": "section", + "value": "Variational Approximation" + }, + { + "attribute": "reference", + "value": "Eq. 7" + } + ] + }, + { + "type": "fact", + "insight": "The RNG design uses only transistors and integrates with traditional circuit components for large-scale sampling systems.", + "content": "Our RNG design uses only transistors and can integrate tightly with other traditional circuit components on a chip to implement a large-scale sampling system. Since there are no exotic components involved that introduce unknown integration barriers, it is straightforward to build a simple physical model to predict how this device utilizes energy.", + "attributes": [ + { + "attribute": "section", + "value": "RNG Design" + }, + { + "attribute": "confidence", + "value": "high" + } + ] + }, + { + "type": "fact", + "insight": "Unit sampling cells implement Boltzmann machine conditional updates as specified in Equation 11.", + "content": "The performance of the device can be understood by analyzing the unit sampling cell that lives on each node of the PGM implemented by the hardware. The function of this cell is to implement the Boltzmann machine conditional update, as given in Eq. 11 in the main text.", + "attributes": [ + { + "attribute": "section", + "value": "Unit Sampling Cell" + }, + { + "attribute": "reference", + "value": "Eq. 11" + } + ] + }, + { + "type": "fact", + "insight": "The sampling cell design uses linear analog circuit to combine neighboring states and model weights, producing control voltage for RNG.", + "content": "There are many possible designs for the sampling cell. The design considered here utilizes a linear analog circuit to combine the neighboring states and model weights, producing a control voltage for an RNG. This RNG then produces a random bit that is biased by a sigmoidal function of the control voltage.", + "attributes": [ + { + "attribute": "section", + "value": "Sampling Cell Design" + }, + { + "attribute": "design_type", + "value": "linear analog circuit" + } + ] + }, + { + "type": "fact", + "insight": "Experimental measurements show the RNG circuit produces random bits at approximately 10MHz rate with ~350aJ energy per bit.", + "content": "We provide experimental measurements of our novel RNG circuitry in the main text, which establish that random bits can be produced at a rate of τ^{-1}_{rng} ≈10MHz using ~350aJ of energy per bit.", + "attributes": [ + { + "attribute": "section", + "value": "Experimental Results" + }, + { + "attribute": "measurement_type", + "value": "energy consumption" + }, + { + "attribute": "confidence", + "value": "experimental" + } + ] + }, + { + "type": "fact", + "insight": "The RNG circuit output voltage waveform wanders randomly between high and low states, with bias being a sigmoidal function of control voltage.", + "content": "Fig. 15 (a) shows an output voltage waveform from the RNG circuit. It wanders randomly between high and low states. Critically, the bias of the RNG circuit (the probability of finding it in the high or low state) is a sigmoidal function of its control voltage, which allows", + "attributes": [ + { + "attribute": "section", + "value": "RNG Circuit Behavior" + }, + { + "attribute": "reference", + "value": "Fig. 15(a)" + }, + { + "attribute": "circuit_behavior", + "value": "sigmoidal bias function" + } + ] + }, + { + "type": "comment", + "insight": "Document includes specific connectivity patterns for graphs of various degrees, with detailed edge mappings for different graph sizes.", + "content": "Pattern Connectivity\nG8 (0,1),(4,1)\nG12 (0,1),(4,1),(9,10)\nG16 (0,1),(4,1),(8,7),(14,9)\nG20 (0,1),(4,1),(3,6),(8,7),(14,9)\nG24 (0,1),(1,2),(4,1),(3,6),(8,7),(14,9)\nTABLE I. Edges (ordered pairs) associated with graphs of various degrees.", + "attributes": [ + { + "attribute": "section", + "value": "Connectivity Patterns" + }, + { + "attribute": "table_reference", + "value": "Table I" + } + ] + }, + { + "type": "comment", + "insight": "The sampling cell design supports initialization and readout operations (get/set state operations) in addition to the main Boltzmann update function.", + "content": "The cell must also support initialization and readout (get/set state operations). A schematic of a unit cell is shown in Fig. 8.", + "attributes": [ + { + "attribute": "section", + "value": "Sampling Cell Operations" + }, + { + "attribute": "reference", + "value": "Fig. 8" + } + ] + }, + { + "type": "fact", + "insight": "The fixed point voltage V∞b is calculated using conductance-weighted sum of bias voltages, with total conductance GΣ being the sum of all individual conductances Gj.", + "content": "V ∞\nb =\nn+2X\nj=1\nGj\nGΣ\nVddyj (D4)\nwhere the total conductanceGΣ is,\nGΣ =\nn+2X\nj=1\nGj (D5)", + "attributes": [ + { + "attribute": "equation", + "value": "D4-D5" + }, + { + "attribute": "type", + "value": "circuit_parameter" + }, + { + "attribute": "voltage", + "value": "V∞b" + } + ] + }, + { + "type": "fact", + "insight": "The RNG bias curve follows a sigmoid function that implements model weights through conductance-weighted input terms and a bias term.", + "content": "P(x i = 1) =σ\n(Vb\nVs\n−ϕ\n)\ninserting Eq. (D4) and expanding the term inside the sigmoid,\nVb\nVs\n−ϕ=\nnX\nj=1\nGj\nGΣ\nVdd\nVs\n(xj ⊕s j) +\n[Gn+1\nGΣ\nVdd\nVs\n−ϕ\n]\n(D7)\nby comparison to the Boltzmann machine conditional, we can see that the first term implements the model weights\n(which can be positive or negative given an appropriate setting of the sign bitsj), and the second term implements\na bias.", + "attributes": [ + { + "attribute": "equation", + "value": "D6-D7" + }, + { + "attribute": "type", + "value": "bias_curve" + }, + { + "attribute": "function", + "value": "sigmoid" + } + ] + }, + { + "type": "fact", + "insight": "Static power consumption in the circuit is proportional to voltage squared and depends on an input-dependent constant γ that represents conductance-weighted sum of yj values.", + "content": "The static power drawn by this circuit can be written in the form,\nP∞ = C\nτbias\nV 2\ndd(1−γ)γ(D8)\nwhere0≤γ≤1is the input-dependent constant,\nγ=\nn+2X\nj=1\nGj\nGΣ\nyj (D9)", + "attributes": [ + { + "attribute": "equation", + "value": "D8-D9" + }, + { + "attribute": "type", + "value": "power_consumption" + }, + { + "attribute": "parameter", + "value": "γ" + } + ] + }, + { + "type": "fact", + "insight": "Energy consumed by the bias circuit is primarily due to static power dissipation, with maximum energy consumption occurring when γ = 1/2.", + "content": "This fixed point must be held while the noise generator relaxes, which means that the energetic cost of the biasing\ncircuit is approximately,\nEbias ≈P∞τrng\n=C τrng\nτbias\nV 2\ndd(1−γ)γ (D10)\nThis is maximized forγ= 1\n2 .\nTo avoid slowing down the sampling machine,τrng\nτbias\n≫1. As such, ignoring the energy spent charging the capacitor\n∼ 1\n2 CV 2\nb will not significantly affect the results, and the approximation made in Eq. (D10) should be accurate. The\nenergy consumed by the bias circuit is primarily due to static power dissipation.", + "attributes": [ + { + "attribute": "equation", + "value": "D10" + }, + { + "attribute": "type", + "value": "energy_consumption" + }, + { + "attribute": "maximum", + "value": "γ=1/2" + } + ] + }, + { + "type": "fact", + "insight": "Communication energy between neighboring cells is determined by wire capacitance and signaling voltage, with energy required being proportional to Cwire × Vsig².", + "content": "In most electronic devices, signals are communicated by charging and discharging wires. Charging a wire\nrequires the energy input,\nEcharge = 1\n2CwireV 2\nsig (D11)\nwhereC wire is the capacitance associated with the wire, which grows with its length, andVsig is the signaling voltage\nlevel.", + "attributes": [ + { + "attribute": "equation", + "value": "D11" + }, + { + "attribute": "type", + "value": "communication_energy" + }, + { + "attribute": "dependency", + "value": "wire_length" + } + ] + }, + { + "type": "fact", + "insight": "Wire capacitance per unit length in the process is approximately 350aF/µm, with total capacitance for node connections calculated using geometric components of connection rules.", + "content": "Given the connectivity patterns shown in table I, it is possible to estimate the total capacitanceCn associated\nwith the wire connecting a node to all of its neighbors,\nCn = 4ηℓ\nX\ni\nq\na2\ni +b 2\ni (D12)\nwhereℓ≈6µmis the sampling cell side length, andη≈350aF/µmis the wire capacitance per unit length in our process, see Fig. 11 (b).ai andb i are thexandycomponents of thei th connection rule, as described in section C.2.", + "attributes": [ + { + "attribute": "equation", + "value": "D12" + }, + { + "attribute": "type", + "value": "capacitance_estimation" + }, + { + "attribute": "parameters", + "value": "ℓ=6µm, η=350aF/µm" + } + ] + }, + { + "type": "fact", + "insight": "Global clock distribution requires signal transmission over long wires with large capacitance, with a simple clock distribution scheme requiring total wire length proportional to the number of rows times row length.", + "content": "Several systems on the chip require signals to be transmitted from some central location out to the individual\nsampling cells. This communication involves sending signals over long wires with a large capacitance, which is\nenergetically expensive. Here, the cost of this global communication will be taken into consideration.\na. Clocking\nAlthough it is possible in principle to implement Gibbs sampling completely asynchronously, in practice, it is more\nefficient to implement standard chromatic Gibbs sampling with a global clock. A global clock requires a signal to\nbe distributed from a central clock circuit to every sampling cell on the chip. This signal distribution is typically\naccomplished using a clock tree, a branching circuit designed to minimize timing inconsistencies between disparate\ncircuit elements.\nTo simplify the analysis, we will consider a simple clock distribution scheme in which the clock is distributed by\nlines that run the entire length of each row in the grid. The total length of the wires used for clock distribution in\nthis scheme is,\nLclock =N L(D13)", + "attributes": [ + { + "attribute": "equation", + "value": "D13" + }, + { + "attribute": "type", + "value": "clock_distribution" + }, + { + "attribute": "scheme", + "value": "row-based" + } + ] + }, + { + "type": "fact", + "insight": "The circuit uses resistors to implement multiply-accumulate operations required by conditional update rules, with resistor conductance tuning needed for specific weight and bias sets.", + "content": "Section D.1 discusses a simple circuit that uses resistors to implement the multiply-accumulate required by the conditional update rule. Key to this is being able to tune the conductance of the resistors to implement specific sets of weights and biases (see Eq. (D7)).", + "attributes": [ + { + "attribute": "section", + "value": "D.1" + }, + { + "attribute": "component", + "value": "resistor circuit" + }, + { + "attribute": "function", + "value": "multiply-accumulate" + } + ] + }, + { + "type": "fact", + "insight": "Memory on-chip storage of model parameters is required for implementing resistor tunability, with writing operations consuming significantly more energy than state maintenance.", + "content": "Practically, implementing this tunability requires that the model parameters be stored in memory somewhere on the chip. Writing to and maintaining these memories costs energy. Writing to the memories uses much more energy than maintaining the state.", + "attributes": [ + { + "attribute": "section", + "value": "D.1" + }, + { + "attribute": "memory_type", + "value": "on-chip" + }, + { + "attribute": "energy_comparison", + "value": "writing >> maintenance" + } + ] + }, + { + "type": "fact", + "insight": "Infrequent memory programming (program once, run many sampling programs) makes maintenance energy cost dominant and minimal compared to sampling cell costs.", + "content": "However, if writes are infrequent (program the device once and then run many sampling programs on it before writing again), then the overall cost of the memory is dominated by maintenance. Luckily, most conventional memories are specifically designed to consume as little energy as possible when not being accessed.", + "attributes": [ + { + "attribute": "section", + "value": "D.1" + }, + { + "attribute": "usage_pattern", + "value": "infrequent writes" + }, + { + "attribute": "cost_dominance", + "value": "maintenance dominates" + } + ] + }, + { + "type": "fact", + "insight": "Memory maintenance energy costs are negligible at system level compared to sampling cell costs and do not significantly affect overall energy outcomes.", + "content": "As such, in practice, the cost of memory maintenance is small compared to the other costs associated with the sampling cells and does not significantly change the outcome shown in Fig. 12.", + "attributes": [ + { + "attribute": "section", + "value": "D.1" + }, + { + "attribute": "impact_level", + "value": "system level" + }, + { + "attribute": "cost_significance", + "value": "negligible" + } + ] + }, + { + "type": "fact", + "insight": "Off-chip communication costs depend heavily on system integration tightness and were analyzed only at chip edge as a conservative lower bound.", + "content": "The cost of this communication depends strongly on the tightness of integration between the two systems and is impossible to reason about at an abstract level. As such, the analysis of communication here (as in Section D3b) was limited to the cost of getting bits out to the edge of our chip, which is a lower bound on the actual cost.", + "attributes": [ + { + "attribute": "section", + "value": "D.2" + }, + { + "attribute": "analysis_scope", + "value": "chip edge only" + }, + { + "attribute": "bound_type", + "value": "lower bound" + } + ] + }, + { + "type": "fact", + "insight": "Detailed PCB-mediated chip communication analysis shows system-level results remain unchanged due to long-running sampling programs.", + "content": "However, we have found that a more detailed analysis, which includes the cost of communication between two chips mediated by a PCB, does not significantly change the results at the system level.", + "attributes": [ + { + "attribute": "section", + "value": "D.2" + }, + { + "attribute": "analysis_detail", + "value": "PCB-mediated" + }, + { + "attribute": "result_impact", + "value": "no significant change" + } + ] + }, + { + "type": "fact", + "insight": "Sampling programs run many iterations before mixing and sending results, causing discrepancy between Esamp and E_init + E_read metrics.", + "content": "The fundamental reason for this is that sampling programs for complex models run for many iterations before mixing and sending the results back to the outside world. This is reflected in the discrepancy betweenEsamp andE init +E read found in section D.4.", + "attributes": [ + { + "attribute": "section", + "value": "D.2" + }, + { + "attribute": "program_behavior", + "value": "many iterations" + }, + { + "attribute": "metric_discrepancy", + "value": "Esamp vs E_init + E_read" + } + ] + }, + { + "type": "fact", + "insight": "Architectural heterogeneity enables sharing of supporting circuitry among sampling cells, dramatically reducing per-cell energy costs.", + "content": "Due to the heterogeneity of our architecture, it is possible to share most of the supporting circuitry among many sampling cells, which dramatically reduces the per-cell cost.", + "attributes": [ + { + "attribute": "section", + "value": "D.3" + }, + { + "attribute": "architecture_type", + "value": "heterogeneous" + }, + { + "attribute": "cost_reduction", + "value": "dramatic per-cell reduction" + } + ] + }, + { + "type": "fact", + "insight": "Supporting circuitry energy costs are insignificant at system level due to architectural sharing capabilities.", + "content": "As such, the energy cost of the supporting circuitry is not significant at the system level.", + "attributes": [ + { + "attribute": "section", + "value": "D.3" + }, + { + "attribute": "cost_level", + "value": "system level" + }, + { + "attribute": "significance", + "value": "insignificant" + } + ] + }, + { + "type": "fact", + "insight": "NVIDIA A100 GPU experiments used Zeus tool for empirical energy measurement and FLOPS-based theoretical estimation.", + "content": "All experiments shown in Fig. 1 in the article were conducted on NVIDIA A100 GPUs. The empirical estimates of energy were conducted by drawing a batch of samples from the model and measuring the GPU energy consumption and time via Zeus [5]. The theoretical energy estimates were derived by taking the number of model FLOPS (via JAX and PyTorch's internal estimators) and plugging them into the NVIDIA GPU specifications (19.5 TFLOPS for Float32 and 400W).", + "attributes": [ + { + "attribute": "section", + "value": "Appendix E" + }, + { + "attribute": "hardware", + "value": "NVIDIA A100 GPU" + }, + { + "attribute": "measurement_method", + "value": "Zeus + FLOPS estimation" + } + ] + }, + { + "type": "fact", + "insight": "Empirical and theoretical GPU energy measurements show good alignment, validating the theoretical estimation approach.", + "content": "The empirical measurements are compared to theoretical estimates for the VAE in Table II, and the empirical measurements show good alignment with the theoretical.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix E" + }, + { + "attribute": "model_type", + "value": "VAE" + }, + { + "attribute": "alignment", + "value": "good empirical-theoretical" + } + ] + }, + { + "type": "fact", + "insight": "GPU energy efficiency data shows empirical measurements consistently higher than theoretical estimates but within reasonable range.", + "content": "FID Empirical Efficiency Theoretical Efficiency\n30.5 6.1×10 −5 2.3×10 −5\n27.4 1.5×10 −4 0.4×10 −4\n17.9 2.5×10 −3 1.7×10 −3", + "attributes": [ + { + "attribute": "section", + "value": "Appendix E" + }, + { + "attribute": "data_type", + "value": "energy efficiency" + }, + { + "attribute": "units", + "value": "joules per sample" + } + ] + }, + { + "type": "comment", + "insight": "The research focuses on relative energy consumption scales rather than state-of-the-art performance, using ResNet and UNet style architectures consistent with literature values.", + "content": "The models were derived from available implementations and are based on ResNet [6] and UNet [7] style architectures. Their FID performance is consistent with published literature values [8–10]. The goal is not to achieve state of the art performance, but to represent the relative scales of energy consumption of the algorithms.", + "attributes": [ + { + "attribute": "section", + "value": "Appendix E" + }, + { + "attribute": "architecture_style", + "value": "ResNet/UNet" + }, + { + "attribute": "research_focus", + "value": "relative energy scales" + } + ] + }, + { + "type": "fact", + "insight": "Diffusion models are substantially less energy-efficient than VAEs due to requiring multiple runs (dozens to thousands of times) to generate a single sample, whereas VAE decoders typically run once.", + "content": "The reader may be surprised to see that the diffusion model is substantially less energy-efficient than the VAE given the relative dominance in image generation. However, two points should be kept in mind. First, while VAE remains a semi-competitive model for these smaller datasets, this quickly breaks down. On larger datasets, a FID performance gap usually exists between diffusion models and VAEs. Second, these diffusion models (based on the original DDPM [2]) have performance that can depend on the number of diffusion time steps. So, not only is the UNet model often larger than a VAE decoder, but it also must be run dozens to thousands of times in order to generate a single sample (thus resulting in multiple orders of magnitude more energy required). Modern improvements, such as distillation [11], may move the diffusion model energy efficiency closer to the VAE's.", + "attributes": [ + { + "attribute": "model_comparison", + "value": "diffusion_models_vs_vae" + }, + { + "attribute": "energy_consumption", + "value": "multiple_orders_of_magnitude" + }, + { + "attribute": "technical_detail", + "value": "UNet_vs_VAE_decoder" + } + ] + }, + { + "type": "opinion", + "insight": "Modern improvements like distillation may help diffusion models achieve energy efficiency closer to VAE levels.", + "content": "Modern improvements, such as distillation [11], may move the diffusion model energy efficiency closer to the VAE's.", + "attributes": [ + { + "attribute": "improvement_suggestion", + "value": "distillation" + }, + { + "attribute": "optimistic_outlook", + "value": "positive" + } + ] + }, + { + "type": "fact", + "insight": "The total correlation penalty gradients can be computed using the same samples used to estimate the gradient of the usual loss in training.", + "content": "The total correlation penalty is a convenient choice in this context because its gradients can be computed using the same samples used to estimate the gradient of the usual loss used in training,∇θLDN .", + "attributes": [ + { + "attribute": "computational_efficiency", + "value": "sample_reuse" + }, + { + "attribute": "optimization", + "value": "gradient_computation" + } + ] + }, + { + "type": "fact", + "insight": "The Adaptive Correlation Penalty (ACP) scheme dynamically adjusts λt based on an estimate of the model's current mixing time using autocorrelation of the Gibbs sampling chain as a proxy.", + "content": "To address this, we employ an Adaptive Correlation Penalty (ACP) scheme that dynamically adjustsλt based on an estimate of the model's current mixing time. We use the autocorrelation of the Gibbs sampling chain,rt yy, as a proxy for mixing, as described in Section H and the main text, Eq. 18.", + "attributes": [ + { + "attribute": "method_name", + "value": "ACP" + }, + { + "attribute": "control_parameter", + "value": "λt" + }, + { + "attribute": "proxy_metric", + "value": "autocorrelation" + } + ] + }, + { + "type": "fact", + "insight": "The ACP algorithm uses a simple layerwise procedure with four steps: estimate current autocorrelation, set minimum lambda value, update lambda based on autocorrelation comparison, and ensure lambda doesn't go below minimum.", + "content": "A simple layerwise procedure is used for this control. The inputs to the algorithm are the initial values ofλt, a target autocorrelation thresholdεACP (e.g.,0.03), an update factorδ ACP (e.g.,0.2) and a lower limitλmin t (e.g.,0.0001).\nAt the end of each training epochm:\n1. Estimate the current autocorrelationat m =r t yy[K]. This estimate can be done by running a longer Gibbs chain periodically and calculating the empirical autocorrelation from the samples.\n2. Setλ′ t =max(λmin t , λ(m) t )to avoid getting stuck at 0.\n3. Updateλ t for the next epoch (m+ 1) based onat m and the previous valueat m−1 (ifm >0):\n•Ifa t m < εACP: The chain mixes sufficiently fast; reduce the penalty slightly.\nλ(m+1) t ←(1−δ ACP)λ′ t\n•Else ifa t m ≥εACP anda t m ≤a t m−1 (orm= 0): Mixing is slow but not worsening (or baseline); keep the penalty strength.\nλ(m+1) t ←λ′ t\n•Else (a t m > εACP anda t m > at m−1): Mixing is slow and worsening; increase the penalty.\nλ(m+1) t ←(1 +δ ACP)λ′ t\n4. If the proposed valueλ(m+1) t < λmin t , then setλ(m+1) t ←0.", + "attributes": [ + { + "attribute": "algorithm_complexity", + "value": "layerwise_procedure" + }, + { + "attribute": "update_logic", + "value": "conditional_adjustments" + } + ] + }, + { + "type": "opinion", + "insight": "The simple feedback mechanism of the ACP algorithm works effectively and is vastly more efficient than manual hyperparameter searches.", + "content": "Our experiments indicate that this simple feedback mechanism works effectively. Whileλt and the autocorrelation at m might exhibit some damped oscillations for several epochs before stabilizing this automated procedure is vastly more efficient than performing manual hyperparameter searches forλt for each of theTmodels.", + "attributes": [ + { + "attribute": "efficiency_claim", + "value": "vastly_more_efficient" + }, + { + "attribute": "automation_benefit", + "value": "reduced_manual_tuning" + } + ] + }, + { + "type": "fact", + "insight": "Training is relatively insensitive to the exact choice of εACP within [0.02,0.1] and δACP within [0.1,0.3], and λmin t within [0.001,0.00001].", + "content": "Training is relatively insensitive to the exact choice ofεACP within a reasonable range (e.g.,[0.02,0.1]) andδ ACP (e.g.,[0.1,0.3]). Assuming that over the course of training theλ t parameter settles around some valueλ ∗ t , one should aim for the lower bound parameterλmin t to be smaller than1 2 λ∗ t , while making sure that the ramp-up time log(λ∗ t )−log(λmin t ) log(1+δACP) remains small. Settings ofλmin t in the range[0.001,0.00001]all produced largely the same result, the only difference being that values on the lower end of that range led to a larger amplitude in oscillations ofλt andat m, but training eventually settled for all values.", + "attributes": [ + { + "attribute": "hyperparameter_robustness", + "value": "wide_range_tolerance" + }, + { + "attribute": "parameter_ranges", + "value": "εACP[0.02,0.1], δACP[0.1,0.3], λmin[0.001,0.00001]" + } + ] + }, + { + "type": "fact", + "insight": "Continuous data can be embedded into binary variables by representing a k-state categorical variable Xi using the sum of k binary variables Zki.", + "content": "In some of our experiments, we needed to embed continuous data into binary variables. We chose to do this by representing ak-state categorical variableXi using the sumkbinary variablesZ k i , Xi = KiX k=1 Z(k) i (G1) whereZ (k) i ∈ {0,1}.", + "attributes": [ + { + "attribute": "embedding_technique", + "value": "binary_representation" + }, + { + "attribute": "variable_type", + "value": "categorical_to_binary" + } + ] + }, + { + "type": "comment", + "insight": "The document appears to be from a machine learning research paper focusing on diffusion models, VAEs, and novel regularization techniques like the Adaptive Correlation Penalty scheme.", + "content": "The reader may be surprised to see that the diffusion model is substantially less energy-efficient than the VAE given the relative dominance in image generation. However, two points should be kept in mind. First, while VAE remains a semi-competitive model for these smaller datasets, this quickly breaks down. On larger datasets, a FID performance gap usually exists between diffusion models and VAEs. Second, these diffusion models (based on the original DDPM [2]) have performance that can depend on the number of diffusion time steps. So, not only is the UNet model often larger than a VAE decoder, but it also must be run dozens to thousands of times in order to generate a single sample (thus resulting in multiple orders of magnitude more energy required). Modern improvements, such as distillation [11], may move the diffusion model energy efficiency closer to the VAE's.", + "attributes": [ + { + "attribute": "document_type", + "value": "research_paper" + }, + { + "attribute": "field", + "value": "machine_learning" + }, + { + "attribute": "focus_area", + "value": "diffusion_models_and_regularization" + } + ] + } + ] + } +}