Book Changelog
This Machine Learning Systems textbook is constantly evolving. This changelog is intended to record all updates and improvements, helping you stay informed about whatβs new and refined.
Automated Changelog
These changelog entries are automatically generated from our development process and should be mostly accurate. They track code changes, content updates, and improvements across the entire book. While the entries are comprehensive, they may occasionally contain minor inaccuracies or overly technical details.
Changelog
2025
π August 18
π Chapters
βββββChapter 2: ML Systems: Improved writing clarity in the ML systems chapter and added a new TikZ figure for better visualization.βββββChapter 6: Data Engineering: Improved clarity of data governance figure and updated labels for the data engineering diagram.βββββChapter 9: Efficient AI: Added a TikZ figure to enhance understanding of neural network architecture.βββββChapter 12: Benchmarking AI: Added a new TikZ figure to illustrate a concept.βββββChapter 18: Robust AI: Clarifies dropoutβs role in uncertainty estimation and elaborates on adversarial example detection.βββββIndex: Updated the βAbout the Bookβ link.
π§βπ» Labs
Seeed XIAO ESP32S3
βββββLab: XIAO Image Classification: Minor text improvements were made to enhance clarity within the Image Classification Lab.βββββLab: XIAO Keyword Spotting: The KWS lab has been updated with new equipment instructions.
Raspberry Pi
βββββLab: Pi Large Language Models: The Ollama lab now uses corrected image paths to display external images in PDF output.
Hands-on Labs
βββββLab: Kits: Updated links to ensure they are accurate.
π August 06
π Chapters
βββββIndex: Improved text wrapping around book cover images for better space utilization.
π August 05
π Frontmatter
βββββAbout: Modernized About the Book section to reflect current organizational structure.βββββChangelog: Updated contentβββββAcknowledgements: Updated contentβββββSocratiQ: Added AI-powered figure caption improvement script.
π Chapters
βββββChapter 1: Introduction: Updated quizzes with new metadata and formatting enhancements.βββββChapter 2: ML Systems: Added quizzes with answers to ML systems chapter and made quiz answer formatting consistent.βββββChapter 3: DL Primer: Enhanced descriptions, clarified key concepts, and added new TikZ figures in chapters 3 through 6. Removed resources sections from chapters, updated section IDs and quiz JSON files, and renamed βConclusionβ sections to βSummaryβ.βββββChapter 4: DNN Architectures: Updated contentβββββChapter 5: AI Workflow: Added quizzes to the AI Workflow chapter with automatic generation from JSON files.βββββChapter 6: Data Engineering: Enhances data engineering section with descriptions and adds new TikZ figures in chapters 3 through 6.βββββChapter 7: AI Frameworks: Added new TikZ figures illustrating framework concepts and enhanced descriptions for improved clarity.βββββChapter 8: AI Training: Added TikZ figures to enhance visual understanding of concepts and improved clarity of explanations.βββββChapter 9: Efficient AI: Added quizzes to the efficient AI chapter with self-check answers and updated quiz formatting.βββββChapter 10: Model Optimizations: Added new TikZ figures to illustrate concepts and improved descriptions for enhanced clarity.βββββChapter 11: AI Acceleration: Added quizzes to the efficient AI chapter.βββββChapter 12: Benchmarking AI: Added new TikZ figures to illustrate concepts within the benchmarking chapter.βββββChapter 13: ML Operations: Enhances descriptions and clarifies key concepts within ML operations.βββββChapter 14: On-Device Learning: Enhances descriptions and clarifies key concepts in On-Device Learning.βββββChapter 15: Security & Privacy: Improved clarity and context of figure captions related to security and privacy concepts.βββββChapter 16: Responsible AI: Improved quiz insertion logic and answer extraction. Added section anchors for self-check answers.βββββChapter 17: Sustainable AI: Updates table caption and column header. Added new TikZ figures in chapter 12.βββββChapter 18: Robust AI: Improved quiz insertion logic and answer extraction, updated some figure captions with added context.βββββChapter 19: AI for Good: Corrections were made to table captions, figure captions, and quiz answers for clarity.βββββChapter 20: Conclusion: Renamed βConclusionβ sections to βSummaryβ and added section anchors for self-check answers.βββββPhD Survival Guide: Quiz answers are now correctly inserted before part blocks when needed.βββββIndex: Added clickable cover image with PDF download functionality and updated book card messaging to early access preview.βββββ404: Updated contentβββββChapter: Emerging Topics: Added placeholder content for emerging topics.βββββChapter: Frontiers: Improved the main page layout by moving the abstract to the beginning, adding a changelog note, and preparing an announcement banner.βββββBest Practices: Added summaries for each part of the book.βββββDesign Principles: Added book part organization.βββββFoundations: Added organization of book parts to improve navigation and understanding.βββββImpact Outlook: The book now includes part summaries which can help readers understand the main points of each section.βββββKits: Updated labs documentation structure, added a lab compatibility matrix, and reordered platforms.βββββLabs: Updated website links to reflect proper navigation between kits and labs.
π§βπ» Labs
Hands-on Labs
βββββLab: Labs Overview: Updated section headers using a script.βββββLab: Lab Setup: Quiz answers are now inserted before part blocks as needed, and section headers have been updated.βββββLab: Nicla Vision: Corrects figure captions to adhere to style guide.βββββLab: Ide Setup: Improved labs documentation with enhanced troubleshooting and platform guides.βββββLab: Kits: Improved labs documentation with enhanced troubleshooting and platform guides.βββββLab: Labs: Improved labs documentation with enhanced troubleshooting and platform guides.βββββLab: Raspi: Updated section ids and headers based on changes to the manager code.βββββLab: Setup And No Code Apps: Updated section headers for improved readability.βββββLab: Xiao Esp32S3: Updated contentβββββLab: Dsp Spectral Features Block: Quiz answers now appear before part blocks when needed.βββββLab: Kws Feature Eng: Updated contentβββββLab: Shared: Updated content
Arduino
βββββLab: Arduino Setup: Updates documentation for XIAO ESP32S3 Sense and improves clarity through minor typo corrections.βββββLab: Arduino Image Classification: Minor typos were corrected for improved clarity.βββββLab: Arduino Object Detection: Corrected minor typos and improved clarity within the lab content.βββββLab: Arduino Keyword Spotting: Updated contentβββββLab: Arduino Motion Classification: Fixed quiz answer insertion logic to appear before part blocks when needed.
Raspberry Pi
βββββLab: Raspberry Pi Setup: Updated section headers using a script for improved consistency.βββββLab: Pi Image Classification: Updated section headers using the script.βββββLab: Pi Object Detection: Updated section headers using a script and changed some section IDs to reflect recent code changes.βββββLab: Pi Large Language Models: Updated contentβββββLab: Pi Vision Language Models: Updated section headers using a script and fixed quiz answer insertion order.
Seeed XIAO ESP32S3
βββββLab: XIAO Setup: Updated section headers for improved readability.βββββLab: XIAO Image Classification: Updated section headers using a script.βββββLab: XIAO Object Detection: Updated section headers using a script to maintain consistency.βββββLab: XIAO Keyword Spotting: Updated section headers using a script.βββββLab: XIAO Motion Classification: Corrected typos, improved wording, and adjusted quiz answer placement within the motion classification lab.
Grove Vision
βββββLab: Grove Vision Ai V2: Updated content
π June 10
π Frontmatter
βββββAbout: Updated SocratiQ page linksβββββSocratiQ: Added documentation for SocratiQ AI learning companion and removed the SocratiQ AI feature.
π Chapters
βββββChapter 1: Introduction: Minor grammatical errors were corrected and the language was refined for improved clarity.βββββChapter 2: ML Systems: Added resource sections to core content and improved text processing in QMD files.βββββChapter 3: DL Primer: Added resource sections to core content, clarified the difference between training and inference, and improved text processing in QMD files for better clarity.βββββChapter 4: DNN Architectures: Refined explanations of deep learning architectures including CNNs, added figures to illustrate data movement patterns, and consolidated footnote definitions for clarity.βββββChapter 5: AI Workflow: Added resource sections to core content, improved text processing in QMD files, and enhanced clarity and consistency.βββββChapter 6: Data Engineering: Added a data pipeline overview diagram and clarified figure references in the text.βββββChapter 7: AI Frameworks: Added resource sections to core content. This update provides additional learning materials beyond the main text.βββββChapter 8: AI Training: Added resource sections to the training content, clarified the activation checkpointing explanation, and improved text processing in QMD files. Figures were also added.βββββChapter 9: Efficient AI: Added resource sections to the core content and clarified the trade-off between efficiency and latency. The scaling laws section was refined for improved clarity.βββββChapter 10: Model Optimizations: Refined model optimization techniques documentation and clarified AutoML and NAS descriptions.βββββChapter 11: AI Acceleration: Improved clarity and accuracy of explanations related to resource allocation in AI accelerators. Added figures and corrected a typo in a matrix multiplication example.βββββChapter 12: Benchmarking AI: Improved clarity and consistency of text related to benchmarking AI.βββββChapter 13: ML Operations: Updated MLOps content for clarity and accuracy. The operations diagram and text were also updated.βββββChapter 14: On-Device Learning: Added resource sections to the core content and clarified explanations of adaptation equations.βββββChapter 15: Security & Privacy: Updated the chapter with expanded discussions on various security vulnerabilities like data poisoning, model theft, and adversarial attacks. Additional content includes a section on trustworthy ML systems and threat mitigation strategies.βββββChapter 16: Responsible AI: Expanded discussions on safety and robustness, fairness, privacy, and data governance in AI. Added a section on design tradeoffs in responsible AI and clarified accountability considerations.βββββChapter 17: Sustainable AI: Added resource sections to the core content and made minor corrections for grammatical errors.βββββChapter 18: Robust AI: Improved clarity and readability of explanations about robust AI techniques.βββββChapter 19: AI for Good: Refined AI for Good content to enhance clarity.βββββIndex: Corrected minor grammatical errors and content inconsistencies.
π§βπ» Labs
Hands-on Labs
βββββLab: Labs Overview: Updated contentβββββLab: Lab Setup: Updated contentβββββLab: Setup And No Code Apps: Improved documentation with corrected latency descriptions and enhanced clarity.
Arduino
βββββLab: Arduino Image Classification: Added Image Classification Lab to the documentation.βββββLab: Arduino Object Detection: Added a new lab focusing on object detection using the Grove Vision AI v2 module.
Seeed XIAO ESP32S3
βββββLab: XIAO Image Classification: Corrected a typo in the image classification lab instructions.
Grove Vision
βββββLab: Grove Vision Ai V2: Added a new lab focused on Grove Vision AI v2.
π May 14
π Chapters
βββββChapter 14: On-Device Learning: On-device learning content was restructured and clarified for improved understanding.
π May 04
π Chapters
βββββChapter 1: Introduction: Updated contentβββββChapter 2: ML Systems: Corrected grammar in a footnote about GDPR/HIPAA compliance.βββββChapter 3: DL Primer: Modified the explanation of dimension ordering for W^L.βββββChapter 4: DNN Architectures: Improved clarity by finding any missing references.βββββChapter 5: AI Workflow: Updated contentβββββChapter 6: Data Engineering: Updated contentβββββChapter 7: AI Frameworks: Updated contentβββββChapter 8: AI Training: Improved label checking for clearer understanding of training data requirements.βββββChapter 9: Efficient AI: Updated contentβββββChapter 10: Model Optimizations: Updated contentβββββChapter 11: AI Acceleration: The discussion on hardware acceleration, specialization, and AI compute primitives has been refined for improved clarity.βββββChapter 12: Benchmarking AI: The benchmarking metrics and power measurements explanations have been clarified.βββββChapter 13: ML Operations: Expanded core MLOps concepts and included additional case studies.βββββChapter 14: On-Device Learning: Added definitions and guidance on on-device learning systems design. Expanded on security concerns, explained privacy in federated learning, and clarified adaptation processes. Included a conclusion, challenges section, tradeoffs summary table, and explorations of on-device learning with limited data and adaptation strategies.βββββChapter 15: Security & Privacy: Improved label checking for accuracy and consistency.βββββChapter 17: Sustainable AI: Improved visual representation of sustainable AI concepts with consolidated TikZ figure styling.βββββChapter 18: Robust AI: Improved label checking for accuracy.βββββChapter 19: AI for Good: Improved accuracy of the PlantVillage Nuru footnote.
π§βπ» Labs
Arduino
βββββLab: Arduino Setup: Instructions were updated for clarity and to correct typos.βββββLab: Arduino Image Classification: Updated image classification lab instructions for improved clarity.βββββLab: Arduino Object Detection: Improved object detection instructions for clarity.βββββLab: Arduino Keyword Spotting: Updated contentβββββLab: Arduino Motion Classification: Updated content
Raspberry Pi
βββββLab: Raspberry Pi Setup: Updated contentβββββLab: Pi Image Classification: Updated contentβββββLab: Pi Object Detection: Updated contentβββββLab: Pi Large Language Models: Updated contentβββββLab: Pi Vision Language Models: The VLM lab guide was restructured for improved clarity.
Seeed XIAO ESP32S3
βββββLab: XIAO Setup: Updated contentβββββLab: XIAO Image Classification: Updated contentβββββLab: XIAO Object Detection: Updated contentβββββLab: XIAO Keyword Spotting: Updated contentβββββLab: XIAO Motion Classification: Updated content
Hands-on Labs
βββββLab: Dsp Spectral Features Block: Updated contentβββββLab: Kws Feature Eng: Updated contentβββββLab: Raspi: Updated contentβββββLab: Xiao Esp32S3: Updated content
π March 25
π Frontmatter
βββββForeword: Updated contentβββββAbout: Updated contentβββββAcknowledgements: Updated contributor list.βββββSocratiQ: Corrected broken links throughout the content.
π Chapters
βββββChapter 1: Introduction: Minor stylistic edits were made to improve readability.βββββChapter 2: ML Systems: Improved footnote consistency and addressing missing references within the ML systems chapter.βββββChapter 3: DL Primer: Improved footnote naming consistency throughout the chapter.βββββChapter 4: DNN Architectures: Corrected hyphenation, improved Markdown styling, fixed broken links, and ensured figure references were accurate.βββββChapter 5: AI Workflow: Added a definition to improve understanding of key concepts within the workflow.βββββChapter 6: Data Engineering: Fixed broken links and made minor text edits to improve clarity.βββββChapter 7: AI Frameworks: Improved figure formatting, ensured consistent footnote naming, and fixed callout formatting for a cleaner presentation of content.βββββChapter 8: AI Training: Improved consistency of footnote naming conventions within the section.βββββChapter 9: Efficient AI: Added a new section on scaling laws and made minor improvements to the existing text.βββββChapter 10: Model Optimizations: Improved clarity of markdown styles and fixed references to figures and tables.βββββChapter 11: AI Acceleration: Improved footnote naming consistency and fixed redundant figure references.βββββChapter 12: Benchmarking AI: Fixed broken links and improved section header clarity.βββββChapter 13: ML Operations: Updated MLOps key components section with narrative structure and restructured core components into groups. Case studies were revised for clarity.βββββChapter 14: On-Device Learning: Fixed broken links within the chapter.βββββChapter 15: Security & Privacy: Fixed broken links.βββββChapter 16: Responsible AI: Fixed broken links within the Responsible AI chapter.βββββChapter 17: Sustainable AI: Added a discussion about Jevonβs paradox and its plot to illustrate the concept.βββββChapter 18: Robust AI: Updated chapter content with new text about robust AI concepts including introductions to poisoning attacks, transient faults and permanent faults. The overview was also improved.βββββChapter 19: AI for Good: Fixed broken links for improved navigational clarity within the chapter.βββββChapter 20: Conclusion: Removed extraneous sections from the conclusion.βββββChapter: Generative Ai: Updated contentβββββChapter: Old Sus Ai: Updated content
π§βπ» Labs
Arduino
βββββLab: Arduino Setup: Corrected broken links within setup instructions.βββββLab: Arduino Image Classification: Improved Markdown styles for better readability.βββββLab: Arduino Keyword Spotting: Updated Markdown styling for improved readability.βββββLab: Arduino Motion Classification: Corrected broken links within the motion classification documentation.
Raspberry Pi
βββββLab: Raspberry Pi Setup: Spelling errors were corrected in the Raspberry Pi setup instructions.βββββLab: Pi Object Detection: Fixed broken links within the Markdown file.βββββLab: Pi Large Language Models: Improved Markdown styles within the document.βββββLab: Pi Vision Language Models: Fixed broken links within the text.
Seeed XIAO ESP32S3
βββββLab: XIAO Image Classification: Spelling mistakes were corrected for improved clarity.βββββLab: XIAO Keyword Spotting: Corrected Markdown styling inconsistencies for improved readability.βββββLab: XIAO Motion Classification: Improved Markdown formatting styles for better readability.
Hands-on Labs
βββββLab: Dsp Spectral Features Block: Improved Markdown style consistency.βββββLab: Kws Feature Eng: Improved Markdown styling for better readability.
π Appendix
βββββPhD Survival Guide: Spelling errors were corrected and all broken links were fixed.
π March 03
π Frontmatter
βββββAbout: Updated contentβββββAcknowledgements: Updated contributors list.βββββSocratiQ: Fixed formatting inconsistencies in callout titles.
π Chapters
βββββChapter 1: Introduction: Fixed formatting issues within callout titles.βββββChapter 2: ML Systems: Corrected markdown formatting issues within the ML Systems chapter.βββββChapter 3: DL Primer: Fixed callout title formatting.βββββChapter 4: DNN Architectures: Fixed formatting issues with callout titles and applied linting fixes to improve QMD file consistency.βββββChapter 5: AI Workflow: Improved text clarity and corrected grammatical errors.βββββChapter 6: Data Engineering: Fixed formatting issues within the data engineering chapter.βββββChapter 7: AI Frameworks: Improved clarity of AI framework descriptions with better formatting and removed redundant information.βββββChapter 8: AI Training: Added descriptions of single and multi GPU systems and removed redundant definitions.βββββChapter 9: Efficient AI: Removed redundant definitions for better clarity.βββββChapter 10: Model Optimizations: Added structured optimization explanations, figures illustrating sparsity and KD, and an LTH + iterative pruning + calibration section. The conclusion was also added.βββββChapter 11: AI Acceleration: Added a section on NVSwitch for multi-GPU setups, included a figure about TPU and updated text with information about models vs. memory bandwidth.βββββChapter 12: Benchmarking AI: Removed an exercise, updated image, and fixed a reference.βββββChapter 13: ML Operations: Fixed formatting of callout titles and addressed QMD linting issues.βββββChapter 14: On-Device Learning: Fixed formatting issues with callouts and improved code readability by removing redundant definitions.βββββChapter 15: Security & Privacy: Fixed formatting inconsistencies in callout titles and improved overall markdown structure.βββββChapter 16: Responsible AI: Fixed formatting issues in callout titles within the Responsible AI chapter.βββββChapter 17: Sustainable AI: Callout title formatting was fixed for improved clarity.βββββChapter 18: Robust AI: Improved formatting and readability of callout titles and overall text.βββββChapter 19: AI for Good: Improved formatting of callout titles within the AI for Good chapter.
π§βπ» Labs
Hands-on Labs
βββββLab: Labs Overview: Updated contentβββββLab: Nicla Vision: Updated contentβββββLab: Kws Feature Eng: Fixed markdown formatting issues in the KWS Feature Engineering documentation.
Arduino
βββββLab: Arduino Setup: Updated contentβββββLab: Arduino Image Classification: Updated the Arduino/Nicla Vision LABS part.βββββLab: Arduino Object Detection: Updated contentβββββLab: Arduino Keyword Spotting: Updated contentβββββLab: Arduino Motion Classification: Linting improved header spacing consistency.
Raspberry Pi
βββββLab: Pi Vision Language Models: Fixed markdown formatting issues in QMD files to ensure proper rendering.
π February 08
π Frontmatter
βββββAcknowledgements: Updated acknowledgements.qmd with contributor information.βββββSocratiQ: Updated content
π February 07
π Frontmatter
βββββAbout: Updated contentβββββChangelog: Updated contentβββββAcknowledgements: Updated contentβββββSocratiQ: The precheck function now only runs on .qmd and .bib files.βββββIndex: Pre-commit checks are now limited to qmd and bib files.
π Chapters
βββββChapter 1: Introduction: The precheck function now only operates on .qmd and .bib files.βββββChapter 2: ML Systems: The precheck script now runs only on qmd and bib files.βββββChapter 3: DL Primer: The precheck script now only runs on .qmd and .bib files.βββββChapter 4: DNN Architectures: Updated contentβββββChapter 5: AI Workflow: The precheck now only runs on .qmd and .bib files.βββββChapter 6: Data Engineering: The precheck now specifically runs on .qmd and .bib files.βββββChapter 7: AI Frameworks: The precheck script now only runs on .qmd and .bib files.βββββChapter 8: AI Training: Improved diagram clarity and formatting.βββββChapter 9: Efficient AI: Added R code for debugging and visualization, addressing feedback regarding existing content.βββββChapter 10: Model Optimizations: The precheck process now specifically targets qmd and bib files.βββββChapter 11: AI Acceleration: Precheck function now focuses specifically on .qmd and .bib files, potentially improving efficiency during document processing.βββββChapter 12: Benchmarking AI: Added new visualizations showcasing power trends in MLPerf benchmarks. The benchmarking challenges chapter now includes a plot demonstrating power ranges and graphs to motivate benchmarking efforts.βββββChapter 13: ML Operations: Precheck now specifically targets qmd and bib files for analysis.βββββChapter 14: On-Device Learning: Updated precheck to focus on qmd and bib files for improved learning resource validation.βββββChapter 15: Security & Privacy: Updated contentβββββChapter 16: Responsible AI: Prechecks now focus exclusively on .qmd and .bib files.βββββChapter 17: Sustainable AI: Prechecks now focus specifically on .qmd and .bib files.βββββChapter 18: Robust AI: The precheck functionality now selectively operates on qmd and bib files.βββββChapter 19: AI for Good: Updated the chapter with improvements to precheck functionality.βββββChapter 20: Conclusion: Updated content
π§βπ» Labs
Hands-on Labs
βββββLab: Labs Overview: The precheck script now only runs on qmd and bib files.βββββLab: Lab Setup: The precheck now only runs on qmd and bib files.βββββLab: Raspi: The precheck script now only runs on qmd and bib files.βββββLab: Dsp Spectral Features Block: The precheck now only runs on .qmd and .bib files.βββββLab: Kws Feature Eng: The precheck now only runs on .qmd and .bib files.βββββLab: Shared: The precheck now only runs on qmd and bib files.
Arduino
βββββLab: Arduino Setup: The precheck now runs only on .qmd and .bib files.βββββLab: Arduino Image Classification: The precheck now only runs on .qmd and .bib files.βββββLab: Arduino Keyword Spotting: The precheck script now focuses on validating .qmd and .bib files only.βββββLab: Arduino Motion Classification: The precheck script now only runs on .qmd and .bib files.
Raspberry Pi
βββββLab: Raspberry Pi Setup: Precheck now focuses solely on .qmd and .bib files.βββββLab: Pi Image Classification: Precheck now only runs on qmd and bib files.βββββLab: Pi Object Detection: The precheck script now only runs on .qmd and .bib files.βββββLab: Pi Large Language Models: The precheck script now only runs on qmd and bib files.βββββLab: Pi Vision Language Models: The precheck now runs only on qmd and bib files.
Seeed XIAO ESP32S3
βββββLab: XIAO Setup: The precheck script now only runs on .qmd and .bib files.βββββLab: XIAO Image Classification: The precheck now only runs on qmd and bib files.βββββLab: XIAO Object Detection: The precheck function now only runs on .qmd and .bib files.βββββLab: XIAO Keyword Spotting: The precheck now focuses solely on .qmd and .bib files.βββββLab: XIAO Motion Classification: The precheck function now only runs on .qmd and .bib files.
π Appendix
βββββPhD Survival Guide: Updated content
π February 02
π Frontmatter
βββββAcknowledgements: Updated content
π Chapters
βββββChapter 1: Introduction: Callout titles throughout the introduction are now presented in a consistent title block format.βββββChapter 2: ML Systems: Callout titles within ###* sections have been updated to a new title block format.βββββChapter 3: DL Primer: Callout titles have been updated to use a title block format for improved visual organization.βββββChapter 4: DNN Architectures: Updated callout titles to a consistent block format.βββββChapter 5: AI Workflow: Callout titles within the AI Workflow section now use a consistent title block format for improved visual clarity.βββββChapter 6: Data Engineering: Updated callout titles to use a title block format for improved readability.βββββChapter 7: AI Frameworks: Improved clarity of TikZ figure usage related to AI frameworks.βββββChapter 8: AI Training: Added several diagrams to enhance understanding of AI training concepts.βββββChapter 9: Efficient AI: Updated callout titles to a title block format and corrected a bibliographic entry.βββββChapter 10: Model Optimizations: Callout titles are now formatted as title blocks.βββββChapter 11: AI Acceleration: Callout titles are now formatted within title blocks for improved visual organization.βββββChapter 12: Benchmarking AI: Improved the learning objectives and benchmark definition. Updated the content with additional figures, case studies, and metrics information.βββββChapter 13: ML Operations: Updated callout titles to use a more consistent title block format.βββββChapter 14: On-Device Learning: Callout titles within the chapter are now formatted using title blocks.βββββChapter 15: Security & Privacy: Callout titles within the chapter are now formatted using title blocks.βββββChapter 16: Responsible AI: Updated callout titles using a title block format for improved visual clarity.βββββChapter 17: Sustainable AI: Callout ###* titles were changed to a title block format for improved visual consistency.βββββChapter 18: Robust AI: Callout titles throughout the chapter have been updated to use a title block format.βββββChapter 19: AI for Good: Callout ###* titles are now formatted using title blocks.
π January 28
π Frontmatter
βββββAcknowledgements: Updated content
π Chapters
βββββChapter 1: Introduction: Removed a redundant case study.βββββChapter 2: ML Systems: Added radar plots to visualize various ML system aspects.βββββChapter 4: DNN Architectures: Wording improvements were made to enhance clarity.βββββChapter 5: AI Workflow: Added a new section explaining prompt engineering techniques for optimizing AI model outputs.βββββChapter 6: Data Engineering: Added new content to the data engineering section with citations and edits to later sections. Keyword research is also underway.βββββChapter 7: AI Frameworks: Added figures to illustrate different types of chips.βββββChapter 8: AI Training: Improved training chapter content with added figures, definitions, explanations about evolution and a conclusion section.βββββChapter 9: Efficient AI: Added learning objectives and made improvements to figures and content.βββββChapter 10: Model Optimizations: Updated contentβββββChapter 11: AI Acceleration: Updated contentβββββChapter 19: AI for Good: Updated learning objectives and added spotlight use cases to demonstrate AI for Good applications.
π§βπ» Labs
Raspberry Pi
βββββLab: Pi Image Classification: Updated contentβββββLab: Pi Object Detection: Updated content
π Appendix
βββββPhD Survival Guide: Added links to helpful resources.
π January 17
π Frontmatter
βββββAbout: Updated contentβββββAcknowledgements: Updated contentβββββSocratiQ: Updated content
π Chapters
βββββChapter 1: Introduction: Addresses feedback regarding content clarity.βββββChapter 2: ML Systems: Updated contentβββββChapter 3: DL Primer: Added explanations of different types of neural networks and clarified the concept of model training.βββββChapter 4: DNN Architectures: Added clarification to parameter storage bound for RNNs.βββββChapter 6: Data Engineering: Updated contentβββββChapter 7: AI Frameworks: Added framework overview, historical context, computational graph section, and updated learning objectives.βββββChapter 12: Benchmarking AI: Updated content
π§βπ» Labs
Raspberry Pi
βββββLab: Pi Large Language Models: Corrected minor copyediting errors.βββββLab: Pi Vision Language Models: Updated content
π January 12
π Frontmatter
βββββAcknowledgements: Added contributors to acknowledgements.
π Chapters
βββββChapter 1: Introduction: Fixed an issue with code rendering that was introduced from PDF enhancements.βββββChapter 2: ML Systems: Added a decision playbook framework and definitions to each section.βββββChapter 5: AI Workflow: Updated contentβββββChapter 6: Data Engineering: Updated data labeling section with fixes and improvements.
π January 11
π Frontmatter
βββββAbout: Updated contentβββββAcknowledgements: Contributors were added to the acknowledgements file.βββββSocratiQ: Updated content
π Chapters
βββββChapter 1: Introduction: Updated the introduction with footnotes.βββββChapter 2: ML Systems: Added a decision playbook framework and provided definitions for each section in the ML Systems chapter.βββββChapter 5: AI Workflow: Updated contentβββββChapter 6: Data Engineering: Updated synthetic data generation methods and clarified explanations about web scraping techniques.
π January 09
π Frontmatter
βββββAcknowledgements: Updated acknowledgements with contributor information.
π Chapters
βββββChapter 1: Introduction: Updated contentβββββChapter 5: AI Workflow: Updated contentβββββChapter 6: Data Engineering: Updated contentβββββChapter 7: AI Frameworks: Updated contentβββββChapter 8: AI Training: Updated contentβββββChapter 11: AI Acceleration: Updated contentβββββChapter 16: Responsible AI: Fixed errors in feedback provided by Bravo.
π January 07
π Frontmatter
βββββForeword: Tweaked wording for improved clarity.βββββAcknowledgements: Updated acknowledgements.qmd with contributors.
π Chapters
βββββChapter 1: Introduction: Improved the explanation of the differences between AI and ML.βββββChapter 3: DL Primer: Added images and code to illustrate the training loop and inference process, including specific examples for training in version 3.5 and inference in version 3.6.βββββChapter 4: DNN Architectures: Added visualization figures and tools to illustrate DNN architectures.
π January 03
π Frontmatter
βββββAcknowledgements: Updated acknowledgements.qmd with contributors.βββββSocratiQ: Updated content
π Chapters
βββββChapter 1: Introduction: Updated contentβββββChapter 2: ML Systems: Updated contentβββββChapter 4: DNN Architectures: Updated contentβββββChapter 6: Data Engineering: Updated contentβββββChapter 20: Conclusion: Updated contentβββββIndex: Fixed mathematical notation errors and improved code examples.
π January 02
π Frontmatter
βββββAcknowledgements: Updated acknowledgements with contributor information.
π Chapters
βββββChapter 4: DNN Architectures: Removed unnecessary commented text.βββββChapter 20: Conclusion: Updated contentβββββIndex: Added HTML tags to enhance the build process.βββββChapter: Generative Ai: Updated content
π January 01
π Frontmatter
βββββForeword: Updated contentβββββAbout: Modified the About section to include Bloomβs Taxonomy concepts and reorganized the content based on feedback.βββββAcknowledgements: Updated contentβββββSocratiQ: Fixed broken links in learning materials and corrected typos.
π Chapters
βββββChapter 1: Introduction: Changed header formats.βββββChapter 2: ML Systems: Expanded Chapter 2 to include a mobile ML section, hybrid ML systems, and an example system.βββββChapter 3: DL Primer: Updated the purpose of the DL Primer chapter.βββββChapter 4: DNN Architectures: Added transformer architecture section with explanations of OG attention and self-attention mechanisms. Updated RNN conclusion and included notes on CNN architectures.βββββChapter 5: AI Workflow: Improved clarity of feedback loops with a new figure and revised explanations.βββββChapter 6: Data Engineering: Updated contentβββββChapter 7: AI Frameworks: Updated the purpose of Chapter 7: AI Frameworks.βββββChapter 8: AI Training: Updated contentβββββChapter 9: Efficient AI: Updated the purpose statement for the chapter.βββββChapter 10: Model Optimizations: Updated contentβββββChapter 11: AI Acceleration: Updated contentβββββChapter 12: Benchmarking AI: Updated purpose statement for benchmarking AI concepts.βββββChapter 13: ML Operations: Updated contentβββββChapter 14: On-Device Learning: Updated the purpose statement for Chapter 14.βββββChapter 15: Security & Privacy: Removed a duplicate case study from the security chapter.βββββChapter 16: Responsible AI: Updated purpose statement.βββββChapter 17: Sustainable AI: Purpose statement was updated.βββββChapter 18: Robust AI: Improved clarity of discussions on BNNs and fault tolerance mechanisms. Refined examples to focus specifically on ML faults and related SDC scenarios.βββββChapter 19: AI for Good: Updated the purpose statement for this chapter.βββββChapter 20: Conclusion: Wording was adjusted in Chapter 20 and a reference to Chapter 4 was included.βββββIndex: Minor updates were made to clarify the bookβs motivation.βββββChapter: Dl Architectures: Updated learning objectives to align with a focus on ML systems and added Colab exercises.
π§βπ» Labs
Hands-on Labs
βββββLab: Labs Overview: Added VLM to the main table.βββββLab: Raspi: Added a new lab related to VLM.
Raspberry Pi
βββββLab: Pi Vision Language Models: Added new Lab - VLMs
2024
π November 19
π Chapters
βββββChapter 15: Security & Privacy: Improved the explanation of power consumption attacks with clearer figures and less repetitive language. Also added a new federated case study.βββββChapter 16: Responsible AI: Improved the presentation of policies discussed in the chapter by adjusting figure placement and refining the figure explanation.βββββChapter 17: Sustainable AI: Added a new figure illustrating the water footprint of AI models and updated the Life Cycle Assessment (LCA) section with new information.βββββChapter 19: AI for Good: The introduction to TinyML was revised to better explain its motivations.βββββAcknowledgements: Updated contentβββββSocratiQ: Added AI podcast
π November 16
π Chapters
βββββChapter 1: Introduction: Improved formatting consistency for definitions.βββββChapter 2: ML Systems: Changed the Introduction to an Overview section.βββββChapter 3: DL Primer: The introduction was renamed to βOverviewββββββChapter 5: AI Workflow: Updated contentβββββChapter 6: Data Engineering: The Introduction section was renamed to Overview.βββββChapter 7: AI Frameworks: The introduction section was changed to an overview section.βββββChapter 8: AI Training: The Introduction was renamed to Overview.βββββChapter 9: Efficient AI: Replaced the Introduction section with an Overview section to provide a more focused introduction to the topic.βββββChapter 10: Model Optimizations: The Introduction was renamed to Overview.βββββChapter 11: AI Acceleration: The introduction section was renamed to βOverviewβ for clarity.βββββChapter 12: Benchmarking AI: Renamed βIntroductionβ to βOverviewβ for conciseness.βββββChapter 13: ML Operations: Improved organization of ML Operations concepts by grouping related topics, streamlining the data management section, and revising the introduction to an overview format.βββββChapter 14: On-Device Learning: The introduction to On-Device Learning has been revised to an overview.βββββChapter 15: Security & Privacy: Changed the Introduction section to an Overview section for better clarity.βββββChapter 16: Responsible AI: Improved clarity of table definitions and reorganized introductory content into an Overview section.βββββChapter 17: Sustainable AI: Renamed the introduction section to βOverviewβ for clarity.βββββChapter 18: Robust AI: The Introduction was changed to an Overview.βββββChapter 19: AI for Good: The Introduction section was renamed to Overview.βββββChapter 20: Conclusion: Revised Introduction to an Overview as there is one main introduction to the material.βββββAbout: Updated contentβββββAcknowledgements: Updated acknowledgements.βββββIndex: Revised preface material for improved organization.βββββContributors: Contributors list was updated.βββββCopyright: Updated contentβββββDedication: Reorganized preface material.
π§βπ» Labs
Arduino
βββββLab: Arduino Setup: Replaced the Introduction section with an Overview section.βββββLab: Arduino Image Classification: The Introduction section was renamed to Overview.βββββLab: Arduino Object Detection: Changed Introduction to Overview as it was the only real introduction present.βββββLab: Arduino Keyword Spotting: The Introduction section was renamed to Overview.βββββLab: Arduino Motion Classification: The Introduction section was renamed to Overview for clarity.
Raspberry Pi
βββββLab: Raspberry Pi Setup: Replaced the Introduction section with an Overview section for clarity.βββββLab: Pi Image Classification: The introduction section was renamed to βOverviewβ.βββββLab: Pi Object Detection: The introduction was renamed to βOverviewβ for clarity.βββββLab: Pi Large Language Models: Changed the Introduction to Overview as there is only one real introduction.
Seeed XIAO ESP32S3
βββββLab: XIAO Setup: Renamed βIntroductionβ to βOverviewβ for improved clarity.βββββLab: XIAO Image Classification: The Introduction section was renamed to Overview for clarity.βββββLab: XIAO Object Detection: The Introduction section was renamed to Overview for better clarity.βββββLab: XIAO Keyword Spotting: The Introduction was renamed to Overview for improved clarity.βββββLab: XIAO Motion Classification: Changed the section title from βIntroductionβ to βOverviewβ for clarity.
Hands-on Labs
βββββLab: Dsp Spectral Features Block: Changed Introduction to Overview as there is one true introduction.βββββLab: Kws Feature Eng: Replaced the βIntroductionβ section with an βOverviewβ section to provide a concise summary of the key concepts.
π November 15
π Chapters
βββββChapter 1: Introduction: Updated introductory text, clarified definitions, added case studies with video/image links, and incorporated feedback to improve the overall flow and content.βββββChapter 2: ML Systems: Revised introduction to encompass a broader scope beyond embedded systems. Updated learning objectives.βββββChapter 3: DL Primer: The introduction was renamed to βOverviewβ and labs were removed from the chapter.βββββChapter 5: AI Workflow: Removed labs content from the workflow chapter.βββββChapter 6: Data Engineering: The Introduction was changed to an Overview and the labs portion was removed from the chapter.βββββChapter 7: AI Frameworks: Removed labs section from the chapter.βββββChapter 8: AI Training: The chapter introduction was renamed to an overview. Labs were removed from this section.βββββChapter 9: Efficient AI: Revised chapter introduction to an overview and removed labs section from the main content.βββββChapter 10: Model Optimizations: Updated content related to model optimizations.βββββChapter 11: AI Acceleration: The introduction was revised to an overview and the labs portion of the chapter was removed.βββββChapter 12: Benchmarking AI: Updated benchmarking content with a new section for energy measurements in historical context, reworked examples, and streamlined descriptions of metrics.βββββChapter 13: ML Operations: Revised Chapter 13 with reorganized topics, a clearer introduction, and updates to the data management section based on feedback.βββββChapter 14: On-Device Learning: The On-Device Learning chapter now provides a clearer distinction between on-device learning and federated learning. Explanations about pruning and IID were improved for better understanding. Lifelong learning advantages are now presented in their own subsection.βββββChapter 15: Security & Privacy: Enhanced the TEE section with additional explanations.βββββChapter 16: Responsible AI: Revised the chapter introduction to an overview and removed lab components.βββββChapter 17: Sustainable AI: Updated content about sustainable AI practices.βββββChapter 18: Robust AI: Removed labs content from the chapter. Updated robustAI content.βββββChapter 19: AI for Good: The chapter introduction was revised to an overview and the labs portion was removed.βββββChapter 20: Conclusion: Revised the Introduction to be an Overview as it is the sole introductory section.βββββAbout: The introduction was moved to the about chapter.βββββAcknowledgements: Updated contentβββββSocratiQ: Updated contentβββββContributors: Updated contentβββββIndex: Removed a link to conventions as it is not currently needed.βββββConventions: Updated content
π§βπ» Labs
Hands-on Labs
βββββLab: Labs Overview: Updated contentβββββLab: Dsp Spectral Features Block: Changed the Introduction section to Overview as there is only one primary introduction.βββββLab: Kws Feature Eng: The Introduction section was renamed to Overview.βββββLab: Labs: Improved documentation and formatting within the labs.βββββLab: Nicla Vision: Corrected formatting of colons in markdown text.βββββLab: Raspi: Fixed inconsistent formatting of text elements.βββββLab: Xiao Esp32S3: Fixed formatting issues with colon usage for better readability.
Arduino
βββββLab: Arduino Setup: Changed Introduction to Overview because there is only one real introduction.βββββLab: Arduino Image Classification: Changed Introduction section to Overview for better clarity.βββββLab: Arduino Object Detection: The introduction was renamed to Overview.βββββLab: Arduino Keyword Spotting: The Introduction section was renamed to Overview.βββββLab: Arduino Motion Classification: The introduction section was renamed to βOverviewβ.
Raspberry Pi
βββββLab: Raspberry Pi Setup: Updated introduction to be more concise and informative.βββββLab: Pi Image Classification: Updated introduction section to be more concise and informative.βββββLab: Pi Object Detection: Changed Introduction to Overview to reflect there is only one introduction section.βββββLab: Pi Large Language Models: Changed the section name from βIntroductionβ to βOverviewβ.
Seeed XIAO ESP32S3
βββββLab: XIAO Setup: Updated the introduction to be more concise and clearly labelled as an overview.βββββLab: XIAO Image Classification: Changed the section heading from βIntroductionβ to βOverviewβ.βββββLab: XIAO Object Detection: The introduction was renamed to βOverviewβ for better clarity.βββββLab: XIAO Keyword Spotting: The introduction section was renamed to Overview for clarity.βββββLab: XIAO Motion Classification: The Introduction section was renamed to Overview.
π September 20
π Chapters
βββββChapter 1: Introduction: Fixed broken figure references.βββββChapter 2: ML Systems: Updated contentβββββChapter 3: DL Primer: Fixed broken links within the chapter.βββββChapter 5: AI Workflow: Updated contentβββββChapter 6: Data Engineering: Fixed inconsistent quotation marks for improved readability.βββββChapter 7: AI Frameworks: Updated contentβββββChapter 8: AI Training: Fixed character formatting issue.βββββChapter 9: Efficient AI: Fixed figure references to ensure accuracy.βββββChapter 10: Model Optimizations: Fixed character formatting inconsistencies.βββββChapter 11: AI Acceleration: Updated contentβββββChapter 12: Benchmarking AI: Removed unnecessary figures from the chapter.βββββChapter 13: ML Operations: Updated contentβββββChapter 14: On-Device Learning: Updated contentβββββChapter 15: Security & Privacy: Updated contentβββββChapter 17: Sustainable AI: Proofreading of the sustainability section corrected typos.βββββChapter 19: AI for Good: Fixed broken figure references.βββββAbout: Updated learning objectivesβββββContributors: Updated content
π§βπ» Labs
Hands-on Labs
βββββLab: Lab Setup: Updated contentβββββLab: Raspi: Updated content
Seeed XIAO ESP32S3
βββββLab: XIAO Setup: Updated contentβββββLab: XIAO Image Classification: Updated contentβββββLab: XIAO Object Detection: Updated contentβββββLab: XIAO Keyword Spotting: Updated contentβββββLab: XIAO Motion Classification: Fixed an image issue.
Raspberry Pi
βββββLab: Raspberry Pi Setup: Corrected character formatting for improved readability.βββββLab: Pi Image Classification: Corrected typographical errors.βββββLab: Pi Object Detection: Corrected typos for improved clarity.
Arduino
βββββLab: Arduino Object Detection: Updated contentβββββLab: Arduino Motion Classification: Updated content
π September 12
π Chapters
βββββChapter 13: ML Operations: Updated contentβββββChapter 17: Sustainable AI: Formatting and stylistic improvements were made to ensure readability.βββββChapter 18: Robust AI: Fixed recommended issues within the Robust AI chapter.βββββChapter 19: AI for Good: Updated contentβββββChapter 20: Conclusion: Updated contentβββββContributors: Updated content
π§βπ» Labs
Raspberry Pi
βββββLab: Pi Image Classification: Corrected a link and typos for improved clarity.βββββLab: Pi Object Detection: Uploaded the Object Detection Lab
π September 06
π Chapters
βββββChapter 16: Responsible AI: Corrected bibliographic information and text formatting.βββββContributors: Updated contributor list.
π September 04
π Chapters
βββββChapter 1: Introduction: Fixed captions to ensure accuracy on even-numbered pages.βββββChapter 2: ML Systems: Updated contentβββββChapter 3: DL Primer: Grammar fixes throughout the chapter.βββββChapter 6: Data Engineering: Updated contentβββββChapter 7: AI Frameworks: Updated contentβββββChapter 8: AI Training: Grammar fixes throughout the chapterβββββChapter 9: Efficient AI: Improved explanations for efficient AI concepts.βββββChapter 10: Model Optimizations: Updated contentβββββChapter 11: AI Acceleration: Improved explanations of AI acceleration techniques.βββββChapter 12: Benchmarking AI: Updated contentβββββChapter 13: ML Operations: Updated contentβββββChapter 14: On-Device Learning: Updated contentβββββChapter 15: Security & Privacy: Grammar fixes were made to improve clarity.βββββChapter 16: Responsible AI: Grammar fixes were made throughout the chapter.βββββChapter 17: Sustainable AI: Grammar fixesβββββChapter 18: Robust AI: Grammar fixesβββββChapter 19: AI for Good: Grammar fixes were applied to improve clarity and readability.βββββChapter 20: Conclusion: Grammar fixesβββββContributors: Updated content
π§βπ» Labs
Arduino
βββββLab: Arduino Image Classification: Updated content
Hands-on Labs
βββββLab: Kws Feature Eng: Updated content
π September 02
π Chapters
βββββChapter 2: ML Systems: Improved sentence flow and clarity.βββββChapter 11: AI Acceleration: Explanations of hardware design principles are now more student-focused.βββββChapter 13: ML Operations: Added a section on model serving within ML Operations.βββββContributors: Updated content
π§βπ» Labs
Raspberry Pi
βββββLab: Raspberry Pi Setup: Updated contentβββββLab: Pi Image Classification: Updated content
π August 29
π Chapters
βββββChapter 13: ML Operations: Updated contentβββββChapter 14: On-Device Learning: On-device learning content was updated based on feedback.βββββContributors: Updated contributors list.βββββIndex: Updated contentβββββTools: Updated content
π§βπ» Labs
Raspberry Pi
βββββLab: Pi Image Classification: Updated content
Hands-on Labs
βββββLab: Labs: Resolved an issue with table merging within the labs content.βββββLab: Kws Feature Eng: Updated content
π August 27
π Chapters
βββββChapter 7: AI Frameworks: Fixed broken links in the hardware acceleration section.βββββChapter 9: Efficient AI: Improved explanations of structure importance methods and corrected an error in figure references.βββββChapter 10: Model Optimizations: Improved explanations of knowledge distillation and adjusted challenges to be more informative.βββββChapter 11: AI Acceleration: Fixed broken links and a duplicate title in the chapter.βββββChapter 12: Benchmarking AI: Updated contentβββββChapter 13: ML Operations: Updated contentβββββChapter 15: Security & Privacy: The Power Attack and Side-Channel Attack sections were edited. Broken links were fixed.βββββChapter 17: Sustainable AI: Fixed broken links within the chapter content.βββββContributors: Updated contentβββββIndex: Minor writing style changes for improved clarity.
π§βπ» Labs
Hands-on Labs
βββββLab: Xiao Esp32S3: Improved the formatting of grid tables for better readability.
π August 22
π Chapters
βββββChapter 11: AI Acceleration: Improved clarity and accuracy of subscript usage examplesβββββChapter 17: Sustainable AI: Added a section on using subscript notation for mathematical expressionsβββββChapter 19: AI for Good: Added usage of subscript formatting.βββββContributors: Updated content
π§βπ» Labs
Raspberry Pi
βββββLab: Raspberry Pi Setup: Updated content
Hands-on Labs
βββββLab: Labs: Updated contentβββββLab: Raspi: Updated content
π August 21
π Chapters
βββββChapter 1: Introduction: Updated contentβββββChapter 2: ML Systems: Updated contentβββββChapter 3: DL Primer: Updated contentβββββChapter 5: AI Workflow: Updated contentβββββChapter 6: Data Engineering: Updated contentβββββChapter 7: AI Frameworks: Updated contentβββββChapter 8: AI Training: Improved table formatting in the chapter.βββββChapter 9: Efficient AI: Updated contentβββββChapter 10: Model Optimizations: Updated contentβββββChapter 11: AI Acceleration: Updated table formatting for improved readability.βββββChapter 12: Benchmarking AI: Updated contentβββββChapter 13: ML Operations: Improved table formatting with striping and hover effects.βββββChapter 14: On-Device Learning: Improved clarity and formatting of on-device learning concepts with a grid table.βββββChapter 15: Security & Privacy: Improved table display with styling updates.βββββChapter 16: Responsible AI: Updated to a grid table for improved presentation.βββββChapter 17: Sustainable AI: Updated contentβββββChapter 18: Robust AI: Improved table formatting with styling enhancements.βββββChapter 19: AI for Good: Updated contentβββββLab: Arduino Image Classification: Removed unnecessary code.βββββAbout: Updated contentβββββContributors: Updated contentβββββDsp Spectral Features Block: Updated contentβββββZoo Datasets: Added Wake Vision dataset to zoo_datasets.qmd.βββββConventions: Updated content
π§βπ» Labs
Raspberry Pi
βββββLab: Lab Setup: Initial setup information added for a Raspberry Pi lab.βββββLab: Raspberry Pi Setup: Initial version of rasPi setup instructions was created.βββββLab: Pi Image Classification: Initial version of rasPi image classification lab provided.βββββLab: Pi Object Detection: Initial version of rasPi object detection lab introduced.βββββLab: Pi Large Language Models: Initial version of rasPiβββββLab: Labs: The initial version of rasPi labs was created.βββββLab: Raspi: Initial version of rasPi content was created.
Seeed XIAO ESP32S3
βββββLab: XIAO Image Classification: Updated contentβββββLab: XIAO Keyword Spotting: Updated content
π August 15
π Chapters
βββββChapter 1: Introduction: Updated contentβββββChapter 2: ML Systems: Updated contentβββββChapter 3: DL Primer: Updated contentβββββChapter 5: AI Workflow: Updated contentβββββChapter 6: Data Engineering: Updated contentβββββChapter 7: AI Frameworks: Addressing typos found in the AI Frameworks section.βββββChapter 8: AI Training: Updated table formatting and made improvements to regularization and hyperparameter search explanations.βββββChapter 9: Efficient AI: Updated contentβββββChapter 10: Model Optimizations: Updated contentβββββChapter 11: AI Acceleration: Updated tables to grid tables for improved visual clarity.βββββChapter 12: Benchmarking AI: Updated contentβββββChapter 13: ML Operations: Improved table presentation with styling enhancements.βββββChapter 14: On-Device Learning: Improved clarity of on-device learning concepts by utilizing a grid table.βββββChapter 15: Security & Privacy: Updated contentβββββChapter 16: Responsible AI: Updated table format to grid style.βββββChapter 17: Sustainable AI: Updated contentβββββChapter 18: Robust AI: Improved table styling with added.stripedand.hoverclasses.βββββChapter 19: AI for Good: Updated contentβββββLab: Arduino Image Classification: Updated contentβββββAbout: Updated contentβββββContributors: Updated contentβββββConventions: Updated content
π§βπ» Labs
Seeed XIAO ESP32S3
βββββLab: XIAO Image Classification: Updated contentβββββLab: XIAO Keyword Spotting: Updated content
π August 07
π Chapters
βββββContributors: Updated contributors list.
π August 06
π Chapters
βββββChapter 1: Introduction: Added HTML + PDF build functionalityβββββChapter 2: ML Systems: Improved the formatting and visual presentation of grid tables in the ML Systems chapter.βββββChapter 3: DL Primer: Corrected broken links to PDFs and videos within the chapter.βββββChapter 5: AI Workflow: Updated tables to use grid formatting for improved readability.βββββChapter 6: Data Engineering: Added a grid table exercise and updated exercises to include βWake Vision Colabβ.βββββChapter 7: AI Frameworks: Made improvements to AI framework descriptions and reduced the focus on federated learning. Added tensor explanations and refined table formatting for improved readability.βββββChapter 8: AI Training: Fixed broken URL links and adjusted table formatting to enhance readability.βββββChapter 9: Efficient AI: Updated table formatting and image references for consistency.βββββChapter 10: Model Optimizations: Improved the formatting of grid tables for better readability.βββββChapter 11: AI Acceleration: Fixed broken URL links and improved formatting consistency for source citations.βββββChapter 12: Benchmarking AI: Improved formatting style for consistency.βββββChapter 13: ML Operations: Improved table formatting consistency and fixed broken links within the content.βββββChapter 14: On-Device Learning: Fixed broken URL links related to PDFs and videos.βββββChapter 15: Security & Privacy: Fixed broken links within the privacy and security section.βββββChapter 16: Responsible AI: Updated sources to be consistent with the text and fixed formatting issues.βββββChapter 17: Sustainable AI: Updated source attribution style for consistency.βββββChapter 18: Robust AI: Improved formatting consistency for tables with markdown, updated source citations and credit style.βββββChapter 19: AI for Good: Improved formatting consistency and added HTML + PDF build functionality.βββββLab: Arduino Image Classification: Fixed image width issues for PDF rendering to ensure accurate visual representation in printed documents.βββββContributors: Updated contentβββββDsp Spectral Features Block: Fixed image width issues to ensure correct PDF rendering.βββββTools: Improved readability of tools tables with left alignment.
π§βπ» Labs
Arduino
βββββLab: Arduino Setup: Corrected image width to ensure proper rendering in PDF format.βββββLab: Arduino Object Detection: Resolved issues affecting PDF rendering of images and fixed broken video links within object detection tutorials.βββββLab: Arduino Keyword Spotting: Fixed issues with image width and URL links to improve rendering and navigation.βββββLab: Arduino Motion Classification: Fixed image rendering issues to ensure proper display of motion classification visualizations in PDFs.
Seeed XIAO ESP32S3
βββββLab: XIAO Setup: Fixed image rendering issues to ensure consistent display across PDF viewers.βββββLab: XIAO Image Classification: Fixed image width issues to ensure proper display and readability in PDF rendering.βββββLab: XIAO Object Detection: Fixed image width issues to ensure proper rendering of object detection visualizations in PDF format.βββββLab: XIAO Keyword Spotting: Fixed image rendering issues to ensure correct display of visual content.βββββLab: XIAO Motion Classification: Fixed image width issues to ensure proper PDF rendering of motion classification diagrams.
Hands-on Labs
βββββLab: Dsp Spectral Features Block: Fixed image width issues for PDF renderingβββββLab: Kws Feature Eng: Fixed image width issues to ensure proper rendering in PDF documents.βββββLab: Nicla Vision: Improved table formatting for better readability.βββββLab: Shared: Improved table readability by aligning text to the left.βββββLab: Xiao Esp32S3: Updated source attribution and improved formatting consistency.
π June 25
π Chapters
βββββChapter 3: DL Primer: Fixed the link to video 3.1.βββββContributors: Updated contributors list.βββββIndex: The banner was added back to the index.
π June 20
π Chapters
βββββChapter 2: ML Systems: Updated contentβββββIndex: The index now includes a banner section with GitHub stars.βββββContributors: Updated contributor list.
π§βπ» Labs
Hands-on Labs
βββββLab: Shared: Fixed broken links within educational content.
π June 19
π Chapters
βββββChapter 1: Introduction: Improved introduction material based on feedback.βββββChapter 2: ML Systems: Improved explanations of ML systems concepts based on feedback from the Data review team.βββββChapter 3: DL Primer: Fixed formatting and typos to improve readability and clarity.βββββChapter 5: AI Workflow: Updated contentβββββChapter 6: Data Engineering: Citation formatting was updated from () to [] for improved consistency.βββββChapter 7: AI Frameworks: Updated contentβββββChapter 8: AI Training: Typographical errors and formatting inconsistencies were corrected.βββββChapter 9: Efficient AI: Added a reference to videos at the relevant link.βββββChapter 10: Model Optimizations: Corrected citation formatting from parentheses to brackets.βββββChapter 11: AI Acceleration: Added a link to Googleβs Edge TPU website.βββββChapter 12: Benchmarking AI: Added a figure illustrating training progress based on MLPerf benchmarks and made minor text updates.βββββChapter 13: ML Operations: Updated contentβββββChapter 14: On-Device Learning: Updated contentβββββChapter 15: Security & Privacy: Updated contentβββββChapter 16: Responsible AI: Updated contentβββββChapter 17: Sustainable AI: Updated contentβββββChapter 18: Robust AI: Fixed citation formatting for improved readability.βββββChapter 19: AI for Good: Updated contentβββββChapter 20: Conclusion: Updated contentβββββLab: Arduino Image Classification: Improved image classification lab integration and added necessary files.βββββForeword: Updated contentβββββAbout: Disabling comments on certain pages.βββββAcknowledgements: Updated contentβββββIndex: The index now includes a banner and links to the GitHub repository.βββββContributors: Updated contentβββββEthics: Updated contentβββββTaxonomy: Updated contentβββββToc: Updated contentβββββLearning Resources: Updated contentβββββDsp Spectral Features Block: Minor change in title.βββββObject Detection Fomo: Updated contentβββββCopyright: Updated contentβββββDedication: Updated contentβββββGenerative Ai: Updated contentβββββLabs: Updated content
π§βπ» Labs
Hands-on Labs
βββββLab: Lab Setup: Added getting started content to the Lab Setup guide.βββββLab: Nicla Vision: Improved the introduction text and added credit for an image.βββββLab: Kws Feature Eng: Updated contentβββββLab: Xiao Esp32S3: Improved introductory text for better clarity.βββββLab: Labs: Updated the overview section of the Labs, and made wording tweaks throughout.βββββLab: Dsp Spectral Features Block: Updated contentβββββLab: Shared: Added Shared Labs overview
Arduino
βββββLab: Arduino Setup: Grammar was corrected and resources were updated.βββββLab: Arduino Object Detection: Improved object detection lab integration within existing course content.βββββLab: Arduino Keyword Spotting: Fixed grammar errors and improved resource links within the Arduino Keyword Spotting lab.βββββLab: Arduino Motion Classification: Improved grammar and syntax within the motion classification lab instructions.
Seeed XIAO ESP32S3
βββββLab: XIAO Setup: Importing SEEED labs and integrating them into existing lab material.βββββLab: XIAO Image Classification: Imported SEEED labs content and integrated it into the image classification section.βββββLab: XIAO Object Detection: Imported materials related to SEEED labs.βββββLab: XIAO Keyword Spotting: Improved readability of lab documentation with grammar corrections and updated link formatting.βββββLab: XIAO Motion Classification: Improved link titles, grammar, and added a link to internal documentation.
π June 11
π Chapters
βββββChapter 2: ML Systems: Improved visual presentation of exercise callouts within the section.βββββChapter 3: DL Primer: Added video callouts and resources at the end of the section.βββββChapter 5: AI Workflow: Added video callouts and resources at the end of the section.βββββChapter 6: Data Engineering: Restructured exercise callouts within the chapter for improved visual presentation.βββββChapter 7: AI Frameworks: Improved the visual presentation of exercise callouts within the section.βββββChapter 8: AI Training: Improved the visual presentation of exercise callouts within the training section.βββββChapter 9: Efficient AI: Added video callouts and end-of-section resources.βββββChapter 10: Model Optimizations: Added video callouts and end of section resources. Improved formatting of exercise callout blocks.βββββChapter 11: AI Acceleration: Added video callouts to enhance section engagement and included end-of-section resources.βββββChapter 12: Benchmarking AI: Improved visual layout of exercise callout blocks.βββββChapter 13: ML Operations: Video callouts were added to the section and exercise callout blocks were reorganized for improved visual appeal.βββββChapter 14: On-Device Learning: Improved the visual presentation of exercise callouts within the section.βββββChapter 15: Security & Privacy: Improved the visual layout of exercise callouts in the section.βββββChapter 16: Responsible AI: Added video callouts and end-of-section resources.βββββChapter 17: Sustainable AI: The chapter now includes video callouts and end-of-section resources. Exercise callout blocks were also reorganized for improved visual appeal.βββββChapter 18: Robust AI: Added video callouts and end-of-section resources. Exercise callout blocks were reorganized for improved visual presentation.βββββChapter 19: AI for Good: Improved visual presentation of exercise callouts within the section.βββββContributors: Updated contentβββββGenerative Ai: Added text about generative AI coming soon.
π June 02
π Chapters
βββββContributors: Updated content
π June 01
π Chapters
βββββChapter 1: Introduction: The introduction section now has improved grammar and readability.βββββChapter 2: ML Systems: Corrected bullet formatting errors and updated slides to ensure proper rendering in PDF.βββββChapter 3: DL Primer: Slides now use a default note style for better PDF rendering. Lab/exercise slides have formatting improvements.βββββChapter 5: AI Workflow: Formatting adjustments were made to labs/exercises/slides for improved PDF rendering.βββββChapter 6: Data Engineering: Fixed an issue with text and URL highlighting in the Data Engineering chapter.βββββChapter 7: AI Frameworks: The βcoming soonβ section now uses bullets and slide formatting was adjusted for better PDF rendering.βββββChapter 8: AI Training: Colab badges are functioning correctly and slides now render well in PDF.βββββChapter 9: Efficient AI: Formatting adjustments were made to labs/exercises/slides for improved PDF rendering.βββββChapter 10: Model Optimizations: Minor formatting updates were made to labs, exercises, and slides for improved PDF rendering.βββββChapter 11: AI Acceleration: Updated slide presentation with bullet points and adjusted formatting for better PDF rendering.βββββChapter 12: Benchmarking AI: Improved formatting of slides and labs/exercises to enhance readability in PDF.βββββChapter 13: ML Operations: Updated coming soon section to have bullets for improved readability.βββββChapter 14: On-Device Learning: Updated slide note formatting for better PDF rendering and improved visual presentation of coming soon sections.βββββChapter 15: Security & Privacy: Improved formatting of slides and labs/exercises to enhance readability in PDF.βββββChapter 16: Responsible AI: The coming soon section was updated with bullets for improved readability. Slides now use a default note style to ensure proper rendering in PDF.βββββChapter 17: Sustainable AI: Slides now use the default note format for improved PDF rendering. Lab/exercise slides were also formatted for consistency.βββββChapter 18: Robust AI: Fixed rendering issues to ensure slides display correctly in PDF format.βββββChapter 19: AI for Good: Updated slides to use a default note style for better PDF rendering and made formatting changes to lab exercises.βββββContributors: Updated contentβββββCase Studies: Fixed rendering issues to ensure all content displays correctly.βββββEthics: Fixed rendering issues to ensure all content displays correctly.βββββGenerative Ai: Fixed rendering issues to ensure content displays correctly.βββββConventions: Improved formatting in labs, exercises, and slides.βββββLabs: Formatting changes were made to slides within the labs exercises.βββββLearning Resources: Updated contentβββββTools: Updated content
π May 26
π Chapters
βββββChapter 1: Introduction: Added a cover image for the introduction chapter and an image related to Markβs article. A reference section was added to the introduction chapter.βββββChapter 2: ML Systems: Added section headers for cross-referencing, updated figure captions and references, corrected grammar, improved clarity of table captions, and changed wording in a few instances.βββββChapter 3: DL Primer: Added section headers for cross-referencing, captions to tables and videos, improved text clarity, and updated resources. Grammar and punctuation were also corrected.βββββChapter 5: AI Workflow: Added section headers for easier cross-referencing and fixed an error in the bib file header.βββββChapter 6: Data Engineering: Added section headers for cross-referencing and captions to all tables. Fixed figure captions and references.βββββChapter 7: AI Frameworks: Improved figure captions and references and added captions to all tables.βββββChapter 8: AI Training: Added section headers for cross-referencing, captions to all tables and short captions for videos. Grammar and punctuation were also checked and fixed.βββββChapter 9: Efficient AI: Added section headers for cross-referencing, corrected figure captions and references, and made punctuation edits.βββββChapter 10: Model Optimizations: Added captions to all tables, short captions for the videos, and added more slides.βββββChapter 11: AI Acceleration: Added short captions for videos and updated hw_acceleration.qmd file with stylistic and link fixes.βββββChapter 12: Benchmarking AI: Updated punctuation, grammar, and styling for improved readability.βββββChapter 13: ML Operations: Added short captions for videos.βββββChapter 14: On-Device Learning: Added captions to all tables and videos, updated the conclusion section, and added exercises.βββββChapter 15: Security & Privacy: Added captions to tables and short captions for videos in the privacy and security section.βββββChapter 16: Responsible AI: Added captions to videos, improved link accuracy, and made minor stylistic changes to enhance readability.βββββChapter 17: Sustainable AI: Added section headers for cross-referencing, improved figure captions and references, and made stylistic changes to improve readability.βββββChapter 18: Robust AI: Added a resources section to the chapter and incorporated feedback from a contributor. Minor text fixes, grammar corrections, punctuation edits, and table formatting adjustments were also made.βββββChapter 19: AI for Good: Added short captions for videos, improved punctuation, and made stylistic changes to text formatting.βββββChapter 20: Conclusion: Made minor grammar and wording improvements to the conclusion section.βββββLab: Arduino Image Classification: Improved punctuation throughout the document for clarity.βββββForeword: Made punctuation edits to improve clarity throughout the foreword.βββββAcknowledgements: Initial draft of the acknowledgements section.βββββContributors: Updated contentβββββIndex: Added a content transparency statement to clarify information sources.βββββCopyright: Updated license fileβββββDedication: Added section headers to aid cross-referencing within the dedication.βββββCase Studies: Added section headers for easier navigation and cross-referencing within case studies.βββββCommunity: Added section headers to facilitate cross-referencing within the community content.βββββEthics: Improved readability by adding section headers for cross-referencing and adjusting the styling of βComing soonβ text.βββββGenerative Ai: Added section headers for improved cross-referencing within the document.βββββLearning Resources: Added section headers for easier navigation and cross-referencing within learning materials.βββββZoo Datasets: Added section headers for easier navigation and cross-referencing within the document.βββββConventions: Corrected punctuation for improved readability.βββββDsp Spectral Features Block: Fixed punctuation errors for improved clarity.βββββKws Feature Eng: Punctuation was corrected.βββββMotion Classify Ad: Punctuation was fixed for improved readability.βββββNiclav Sys: Fixed punctuation inconsistencies for improved readability.βββββToc: Corrected punctuation inconsistencies throughout the document.βββββLabs: Updated βtinyMLβ references to βTinyMLβ.βββββEmbedded Ml: Set collapse=false to control chapter visibility.βββββEmbedded Sys: Added more slides to enhance visual learning.
π March 21
π Chapters
βββββChapter 3: DL Primer: Added a Resources section to each part of the DL Primer with introductory text and collapsed functionality. Additional slides were also incorporated.βββββChapter 5: AI Workflow: The Resources section now includes introductory text for each part and can be collapsed. Slides have been moved to the end of the page.βββββChapter 6: Data Engineering: Added a βResourcesβ section to all QMDs with collapsible intro text for each part.βββββChapter 7: AI Frameworks: Added a section with introductory text and enabled collapsible sections within the Resources.βββββChapter 8: AI Training: Added a Resources section at the end of the chapter with introductory text and enabled collapsing functionality for better organization.βββββChapter 9: Efficient AI: Added more slides and an βResourcesβ section with introductory text that can be collapsed.βββββChapter 10: Model Optimizations: Added an empty βResourcesβ section at the end of each QMD file to allow for future material additions.βββββChapter 11: AI Acceleration: Added introductory text for each section within the Resources part and enabled collapsible sections.βββββChapter 12: Benchmarking AI: Added an βResourcesβ section at the end of all QMDs with intro text and enabled collapsing.βββββChapter 13: ML Operations: Added a Resources section at the end of each QMD with introductory text for each part and enabled collapsing.βββββChapter 14: On-Device Learning: Resources section added to the end of all QMDs, with collapsible intro text and space for learning materials.βββββChapter 15: Security & Privacy: Added a βResourcesβ section with intro text and enabled collapsing at the end of all QMDs.βββββChapter 16: Responsible AI: Added a Resources section to the end of the Responsible AI chapter with collapsible sections for each resource category.βββββChapter 17: Sustainable AI: Added an empty βResourcesβ section to the end of the QMD with headers.βββββChapter 19: AI for Good: Added an empty βResourcesβ section to the end of the QMD with headers and updated the page by moving slides to the end.βββββContributors: Updated contentβββββLabs: Added a Labs QMD file to provide information and recognition for Marceloβs contributions.βββββEmbedded Sys: Added a Resources section to QMD files with intro text for each part and enabled collapsing.βββββEmbedded Ml: Added a Resources section with introductory text and collapsible features to each part.
π March 13
π Chapters
βββββContributors: Updated content
π March 12
π Chapters
βββββChapter 1: Introduction: Updated contentβββββChapter 3: DL Primer: Added more slides.βββββChapter 5: AI Workflow: Added more slides and fixed notes from last week.βββββChapter 6: Data Engineering: Updated the Data Engineering chapter with Colab notebooks, added more slides, and included a web scraping exercise in both the subsection and as a separate Exercises part.βββββChapter 7: AI Frameworks: Added Colab notebooks to provide hands-on experience with AI frameworks covered in the chapter.βββββChapter 8: AI Training: Improved the visual presentation of AI training content.βββββChapter 9: Efficient AI: Improved correctness of non-ASCII character handling scripts.βββββChapter 10: Model Optimizations: Updated contentβββββChapter 11: AI Acceleration: Removed a figure reference and mermaid section from the text.βββββChapter 12: Benchmarking AI: Added additional slides to enhance the presentation of benchmarking concepts.βββββChapter 13: ML Operations: Added more slides about ML Operations.βββββChapter 14: On-Device Learning: Added more slides.βββββChapter 15: Security & Privacy: Added more slides to enhance visual learning.βββββChapter 16: Responsible AI: Improved visual styling of slides for better presentationβββββChapter 17: Sustainable AI: Added more slides to enhance presentation coverage of sustainable AI topics.βββββChapter 19: AI for Good: Added more slides.βββββAcknowledgements: Updated contentβββββContributors: Updated contentβββββNiclav Sys: Fixed incorrect links.βββββEmbedded Ml: Added slides with arrow capabilities, custom callouts, and more content.βββββEmbedded Sys: Added more slides about embedded systems concepts.βββββTools: Non-ASCII checker scripts were added and existing ones were fixed.
π February 03
π Chapters
βββββChapter 3: DL Primer: Fixed video rendering issues.βββββChapter 11: AI Acceleration: Fixed video rendering for improved visual demonstration of AI acceleration techniques.βββββChapter 12: Benchmarking AI: Improved the visual consistency of the benchmarking sectionβs list items.βββββChapter 13: ML Operations: Added an MCU example for smartwatch implementation and included a relevant reference.βββββChapter 14: On-Device Learning: Fixed rendering of itemized lists for improved readability.βββββChapter 15: Security & Privacy: Improved clarity and added hyperlinking to relevant sections for GDPR and CCPA compliance guidelines.βββββChapter 17: Sustainable AI: Improved formatting of list items and cited a reference for an OECD blueprint paper.βββββChapter 19: AI for Good: Fixed video rendering issues and resolved YouTube shortened URL resolution problems.βββββContributors: Updated content
π February 02
π Chapters
βββββChapter 3: DL Primer: Updated image format for PDF builds to PNG.βββββChapter 6: Data Engineering: Added a web scraping exercise using Google Colab.βββββChapter 8: AI Training: Updated contentβββββChapter 10: Model Optimizations: Improved illustration of sparsity matrix filter.βββββChapter 11: AI Acceleration: Fixed several broken image references within the chapter.βββββChapter 12: Benchmarking AI: Fixed rendering of references within the benchmarking chapter.βββββChapter 13: ML Operations: Fixed rendering issues with a figure.βββββChapter 14: On-Device Learning: Updated formatting and removed a broken image reference.βββββChapter 15: Security & Privacy: Security section content now renders correctly with fixed image references and video URLs.βββββChapter 16: Responsible AI: Fixed an issue with citations using the β@β symbol for consistency.βββββChapter 17: Sustainable AI: Fixed several broken image references and links within the chapter.βββββChapter 19: AI for Good: Fixed broken image references to ensure all figures are displayed correctly.βββββContributors: Updated contentβββββEmbedded Sys: Bibtex references are now updated automatically.βββββEmbedded Ml: Updated content
π January 02
π Chapters
βββββChapter 7: AI Frameworks: Minor syntax errors were corrected in callout-tip elements.βββββContributors: Updated contentβββββNiclav Sys: Corrected a typo in the instructions for installing the OpenMV IDE.
2023
π December 19
π Chapters
βββββChapter 10: Model Optimizations: Added figures to illustrate model optimization concepts and corrected formatting errors.βββββContributors: Updated contributor list
π December 18
π Chapters
βββββChapter 7: AI Frameworks: Updated Colab notebooks for AI frameworks examples.βββββChapter 10: Model Optimizations: Updated contentβββββChapter 12: Benchmarking AI: Content about benchmarking has been moved to a new section within the benchmarks/leaderboards area. The display of references has also been improved.βββββChapter 17: Sustainable AI: Improved wording about power draw and fixed a citation issue.βββββLearning Resources: Improved readability by removing line wraps.
π December 13
π Chapters
βββββChapter 7: AI Frameworks: Colab notebooks for frameworks were updated.βββββChapter 8: AI Training: Updated contentβββββChapter 9: Efficient AI: Fixed a broken URL link.βββββChapter 10: Model Optimizations: Updated a missing reference to an attention paper for further reading.βββββChapter 12: Benchmarking AI: Updated contentβββββLearning Resources: Removed an invalid learning resource.βββββIndex: Added a section on how to cite the book to the preface.
π December 12
π Chapters
βββββChapter 1: Introduction: The introduction paragraph was updated to explain scholarly paper links.βββββChapter 3: DL Primer: Removed the DL primer activation function explanation and moved the computation graph discussion to the training section.βββββChapter 5: AI Workflow: The name βtinyMLβ was updated to be consistently written as βTinyMLβ throughout the workflow documentation.βββββChapter 6: Data Engineering: Updated βtinyMLβ instances to be consistently written as βTinyMLβ.βββββChapter 7: AI Frameworks: Ensured consistent terminology by updating βtinyMLβ to βTinyMLβ.βββββChapter 8: AI Training: Removed the discussion of activation function from the deep learning primer and moved the computation graph description to the training section.βββββChapter 10: Model Optimizations: Minor language edits for consistency.βββββChapter 11: AI Acceleration: Consistently used βTinyMLβ throughout the text.βββββChapter 12: Benchmarking AI: Updated βtinyMLβ terminology to be consistently written as βTinyMLβ.βββββChapter 14: On-Device Learning: Updated contentβββββChapter 16: Responsible AI: Ensured consistent terminology by changing βtinyMLβ to βTinyMLβ.βββββChapter 18: Robust AI: Updated contentβββββLab: Arduino Image Classification: Updated contentβββββIndex: Added a βHow to Cite This Bookβ section to the preface.βββββGenerative Ai: Updated contentβββββEmbedded Ml: Updated contentβββββEmbedded Sys: Updated terminology to be consistent with current industry standards.βββββKws Nicla: Updated terminology for consistency.βββββTools: Ensured consistent terminology by replacing instances of βtinyMLβ with βTinyMLβ.βββββZoo Datasets: Ensured consistent terminology by changing βtinyMLβ to βTinyMLβ throughout.
π December 11
π Chapters
βββββChapter 1: Introduction: The introductory paragraph now explains the purpose of links to scholarly papers.βββββChapter 3: DL Primer: Updated contentβββββChapter 5: AI Workflow: Updated contentβββββChapter 6: Data Engineering: Updated contentβββββChapter 7: AI Frameworks: Updated contentβββββChapter 8: AI Training: Moved computation graph implementation to training section.βββββChapter 9: Efficient AI: Added references to mentioned datasets and ResNet-SE and ResNeXt papers in the efficient AI chapter.βββββChapter 10: Model Optimizations: Removed duplicate information about the lottery ticket hypothesis.βββββChapter 11: AI Acceleration: Added references for Machine Learning/Reinforcement Learning algorithms in hardware design applications such as architecture design exploration, floorplanning, and logic synthesis.βββββChapter 12: Benchmarking AI: Updated contentβββββChapter 13: ML Operations: Updated contentβββββChapter 14: On-Device Learning: Updated contentβββββChapter 15: Security & Privacy: Updated contentβββββChapter 16: Responsible AI: Updated contentβββββChapter 17: Sustainable AI: Updated contentβββββChapter 18: Robust AI: Updated contentβββββChapter 19: AI for Good: Updated contentβββββLab: Arduino Image Classification: Organized image files by type to enhance clarity.βββββGenerative Ai: Updated contentβββββEmbedded Ml: Organized images into subfolders based on file type for easier navigation.βββββEmbedded Sys: The embedded systems documentation now uses consistent terminology throughout and includes separate reference files for each chapter.βββββKws Nicla: Updated contentβββββTools: Updated terminology to be consistent throughout.βββββZoo Datasets: Updated language consistency regarding TinyML.βββββIndex: Consistency was improved by updating references to βTinyMLβ throughout the text.βββββDsp Spectral Features Block: Updated contentβββββKws Feature Eng: Updated contentβββββMotion Classify Ad: Updated contentβββββNiclav Sys: Updated contentβββββObject Detection Fomo: Updated contentβββββContributors: Updated content
π December 10
π Chapters
βββββChapter 1: Introduction: Updated contentβββββChapter 3: DL Primer: Updated contentβββββChapter 5: AI Workflow: Updated contentβββββChapter 6: Data Engineering: Updated contentβββββChapter 7: AI Frameworks: Updated contentβββββChapter 8: AI Training: Updated contentβββββChapter 9: Efficient AI: Updated contentβββββChapter 10: Model Optimizations: Updated contentβββββChapter 11: AI Acceleration: Updated contentβββββChapter 12: Benchmarking AI: Updated contentβββββChapter 13: ML Operations: Updated contentβββββChapter 14: On-Device Learning: Updated contentβββββChapter 15: Security & Privacy: Updated contentβββββChapter 16: Responsible AI: Updated contentβββββChapter 17: Sustainable AI: Updated contentβββββChapter 19: AI for Good: Updated contentβββββLab: Arduino Image Classification: Updated contentβββββContributors: Updated contentβββββIndex: Fixed broken links and updated contact information.βββββDsp Spectral Features Block: Updated contentβββββEmbedded Ml: Updated contentβββββEmbedded Sys: Updated contentβββββGenerative Ai: Updated contentβββββKws Feature Eng: Updated contentβββββKws Nicla: Updated contentβββββMotion Classify Ad: Updated contentβββββNiclav Sys: Updated contentβββββObject Detection Fomo: Updated content
π December 09
π Chapters
βββββChapter 6: Data Engineering: Minor improvements were made to references within the chapter.βββββChapter 11: AI Acceleration: Added references and fixes related to CPU and GPU acceleration techniques.βββββContributors: Updated the list of contributors to the project.
π December 08
π Chapters
βββββChapter 5: AI Workflow: Fixed figure reference for improved visual clarity.βββββChapter 6: Data Engineering: Updated contentβββββChapter 7: AI Frameworks: Added exercises to the AI Frameworks chapter and included new figures illustrating key concepts.βββββChapter 8: AI Training: Updated contentβββββChapter 9: Efficient AI: Fixed spelling errors throughout the chapter.βββββChapter 10: Model Optimizations: Changed the list format from effective to bulleted.βββββChapter 17: Sustainable AI: Added a reference to nuclear data centers and made minor formatting updates to sustainable_ai.qmd.βββββContributors: Updated contentβββββMotion Classif Anomaly Detect: Including exercises on FrameworkβββββMotion Classify Ad: Added exercises on FrameworkβββββEmbedded Ml: Fixed figure reference for improved visual clarity.
π December 06
π Chapters
βββββChapter 1: Introduction: Updated contentβββββChapter 3: DL Primer: Added exercises focusing on deep learning frameworks.βββββChapter 5: AI Workflow: Updated contentβββββChapter 6: Data Engineering: Added figures to illustrate embedded_ai, ai_workflow, and data engineering concepts.βββββChapter 7: AI Frameworks: Fixed markdown formatting issues.βββββChapter 8: AI Training: Updated the training parallelization section, improved the optimizations section, added details to activation functions, and made weight initialization connections clearer.βββββChapter 9: Efficient AI: Added visualizations to enhance understanding of concepts.βββββChapter 10: Model Optimizations: Updated contentβββββChapter 11: AI Acceleration: Updated contentβββββChapter 12: Benchmarking AI: Updated contentβββββChapter 13: ML Operations: Updated contentβββββChapter 14: On-Device Learning: Corrected a typo to ensure consistency in terminology.βββββChapter 15: Security & Privacy: Updated contentβββββChapter 16: Responsible AI: Updated sections on autonomous systems, AI safety and value alignment, interpretable models, bias and privacy. Added a cover image, learning objectives, and revised the introduction.βββββChapter 17: Sustainable AI: Added citations and images to the chapter on Sustainable AI. Content was also updated with a first draft of the chapter.βββββChapter 19: AI for Good: Updated contentβββββLab: Arduino Image Classification: Exercises now include cover images for improved visual appeal and context.βββββAcknowledgements: Updated contentβββββEmbedded Ml: Added figures to illustrate embedded ML concepts related to cloud ML.βββββIndex: Corrected typos in the index file.βββββContributors: Updated contentβββββMotion Classif Anomaly Detect: Added new exercises with cover images to enhance visual appeal and engagement.βββββDsp Spectral Features Block: Updated contentβββββKws Feature Eng: Exercises now include cover images to provide visual context.βββββKws Nicla: Added cover images to exercises.βββββLearning Resources: Exercises now include cover images.βββββNiclav Sys: Added exercises covering frameworks and deep learning primer concepts.βββββObject Detection Fomo: Added cover images to exercises for enhanced visual appeal and engagement.βββββZoo Models: Updated contentβββββZoo Datasets: Updated contentβββββTools: Updated contentβββββTest: Updated contentβββββGenerative Ai: Updated contentβββββEmbedded Sys: Updated contentβββββCopyright: Updated contentβββββCommunity: Updated contentβββββCase Studies: Updated content
π December 01
π Chapters
βββββChapter 6: Data Engineering: Updated figures and tables within the data engineering section for improved clarity.βββββChapter 8: AI Training: Improved the clarity and accuracy of the hyperparameter section.βββββChapter 15: Security & Privacy: Minor updates were made to enhance clarity and accuracy of information regarding privacy and security concepts.βββββContributors: Updated content
π November 30
π Chapters
βββββChapter 6: Data Engineering: Updated image descriptions with copyright attribution and added five visuals to enhance learning.βββββChapter 8: AI Training: The algorithms section was expanded with additional information and references.βββββChapter 9: Efficient AI: Made table formatting consistent.βββββChapter 10: Model Optimizations: Updated contentβββββChapter 11: AI Acceleration: Updated contentβββββChapter 13: ML Operations: Removed duplicate references to ensure clarity and accuracy.βββββChapter 14: On-Device Learning: Improved the visual consistency of tables.βββββChapter 15: Security & Privacy: Updated the chapter with corrections to references and formatting.βββββContributors: Updated contentβββββIndex: Updated content
π November 22
π Chapters
βββββChapter 8: AI Training: Updated backpropagation explanation.βββββChapter 13: ML Operations: Incorporated feedback to improve clarity and accuracy of ML Operations content.βββββChapter 15: Security & Privacy: Added a cover image, learning objectives, and a draft chapter on security and privacy.βββββContributors: Updated contentβββββEmbedded Sys: Added images to illustrate the difference between microcontrollers and microprocessors.
π November 17
π Chapters
βββββChapter 5: AI Workflow: Updated text to align with an image illustrating the traditional machine learning workflow and added a corresponding image to the chapter.βββββChapter 8: AI Training: Added training data content and an overview of neural networks.βββββChapter 11: AI Acceleration: Added a link to Neuromorphic Computing within the chapter.βββββChapter 12: Benchmarking AI: Added a section link to Neuromorphic Computing.βββββChapter 13: ML Operations: Updated acronyms used throughout Chapter 13.βββββContributors: Updated contentβββββIndex: The introduction was made more general.
π November 15
π Chapters
βββββChapter 8: AI Training: Added training data content, an introduction to neural networks, and placeholders for additional sections.βββββChapter 11: AI Acceleration: Fixed spelling errors and improved figure accuracy.βββββChapter 13: ML Operations: Added an overview paragraph and a page dedicated to AIOps.βββββContributors: Updated content
π November 12
π Chapters
βββββContributors: Fixed a broken link to a book listed in the contributors.
π November 10
π Chapters
βββββChapter 12: Benchmarking AI: Updated contentβββββContributors: Updated contributor list.
π November 09
π Chapters
βββββChapter 9: Efficient AI: Updated contentβββββChapter 11: AI Acceleration: Added sections on Software for AI hardware and Benchmarking AI Hardware. Also included a co-design section with references. Content was added on emerging technologies, an introduction to hardware accelerators, types of hardware accelerators, and background information.βββββChapter 13: ML Operations: Updated the image for better visualization.βββββChapter 14: On-Device Learning: Updated the advantages and limitations section based on feedback and revised the transfer learning section to address comments.βββββContributors: Updated contentβββββDsp Spectral Features Block: Updated content
π November 07
π Chapters
βββββChapter 9: Efficient AI: Added learning objectives section to guide student understanding.βββββChapter 11: AI Acceleration: Added sections on software for AI hardware, benchmarking AI hardware, co-design considerations, and emerging technologies in AI acceleration. Included background information, types of hardware accelerators, and references.βββββChapter 13: ML Operations: Updated image for improved visual clarity.βββββChapter 14: On-Device Learning: Updated the advantages and limitations section of on-device learning with additional information based on feedback.βββββChapter 19: AI for Good: Added a medical example to illustrate AI applications within the βAI for Goodβ chapter.βββββContributors: Updated contentβββββDsp Spectral Features Block: Updated content
π November 03
π Chapters
βββββObject Detection Fomo: Added Exercise Motion/Anomaly DetectionβββββContributors: Updated contributor list.
π November 02
π Chapters
βββββChapter 5: AI Workflow: Updated contentβββββChapter 6: Data Engineering: Updated contentβββββChapter 10: Model Optimizations: Added an overview paragraph about the chapter.βββββChapter 11: AI Acceleration: Updated the figure illustrating AI acceleration concepts.βββββChapter 13: ML Operations: Updated contentβββββChapter 14: On-Device Learning: Added learning objectives and citation links to the Transfer Learning section.βββββChapter 17: Sustainable AI: Added a cover image for Chapter 17.βββββChapter 19: AI for Good: Updated cover image.βββββContributors: Updated content
π October 31
π Chapters
βββββChapter 3: DL Primer: Notes within the chapter no longer use collapsible sections.βββββChapter 5: AI Workflow: Notes section no longer collapses by default.βββββChapter 6: Data Engineering: Improved readability by removing unnecessary collapsing on notes.βββββChapter 7: AI Frameworks: Improved note section readability by removing collapsed sections.βββββChapter 8: AI Training: Notes within the chapter no longer have an automatic collapse.βββββChapter 9: Efficient AI: Notes within the chapter no longer collapse by default.βββββChapter 10: Model Optimizations: Fixed mathematical notation errors and improved code examples for model optimizations.βββββChapter 11: AI Acceleration: Updated contentβββββChapter 12: Benchmarking AI: Minor formatting adjustments were made to improve readability of notes.βββββChapter 13: ML Operations: Improved readability by removing unnecessary collapse functionality on notes.βββββChapter 14: On-Device Learning: Improved readability by removing the collapse functionality from note sections.βββββChapter 15: Security & Privacy: Updated contentβββββChapter 16: Responsible AI: Removed collapse on notesβββββChapter 18: Robust AI: Removed unnecessary collapse on notes.βββββChapter 19: AI for Good: Removed collapsing on notes for improved readability.βββββContributors: Updated contentβββββCase Studies: Updated contentβββββEmbedded Ml: Removed collapsing functionality on notes.βββββEmbedded Sys: Removed the collapsible feature from notes sections.βββββEthics: Updated contentβββββGenerative Ai: Removed collapse functionality from notes section.
π October 30
π Chapters
βββββChapter 3: DL Primer: Added DALLE3 figures to enhance visual understanding of concepts.βββββChapter 5: AI Workflow: Added DALLE3 figures to enhance visual understanding of concepts.βββββChapter 6: Data Engineering: Added DALLE3 figures to enhance visual learning.βββββChapter 7: AI Frameworks: Updated framework cover image.βββββChapter 8: AI Training: Notes within collapsed sections are now visible.βββββChapter 9: Efficient AI: Removed unnecessary collapse from notes.βββββChapter 10: Model Optimizations: Updated contentβββββChapter 11: AI Acceleration: Updated contentβββββChapter 12: Benchmarking AI: Updated chapter structure with a conclusion, learning objectives, and introduction material. Added images related to benchmarking and references.βββββChapter 13: ML Operations: Improved clarity of notes by removing collapsed sections.βββββChapter 14: On-Device Learning: Notes no longer collapse automatically.βββββChapter 15: Security & Privacy: Updated contentβββββChapter 16: Responsible AI: Removed collapse functionality from notes sectionβββββChapter 18: Robust AI: Removed the collapsing functionality on notes.βββββChapter 19: AI for Good: Notes section no longer uses collapsible elements.βββββCase Studies: Notes no longer collapse automatically.βββββEmbedded Ml: Added visual explanations using DALLE3 figures to enhance understanding of some concepts.βββββEmbedded Sys: Added visuals of DALLE3 figures to several chapters.βββββEthics: Notes no longer collapse automatically.βββββGenerative Ai: The collapse functionality has been removed from notes.βββββContributors: Updated contentβββββKws Nicla: Updated content
π October 29
π Chapters
βββββChapter 7: AI Frameworks: Learning objectives were updated.βββββChapter 9: Efficient AI: Updated contentβββββChapter 12: Benchmarking AI: Updated contentβββββLab: Arduino Image Classification: Added Hands-On ExercisesβββββContributors: Updated contentβββββKws Nicla: Added Hands-On Exercises to enhance practical understanding.βββββEmbedded Ml: Added Hands-On ExercisesβββββEmbedded Ml Exercise: Added Hands-On ExercisesβββββEmbedded Sys: Added Hands-On ExercisesβββββEmbedded Sys Exercise: Added Hands-On ExercisesβββββKws Feature Eng: Added Hands-On Exercises to enhance practical understanding of concepts.βββββNiclav Sys: Added Hands-On ExercisesβββββObject Detection Fomo: Added Hands-On ExercisesβββββCommunity: Added a link to the TinyML Edu webpage.
π October 24
π Chapters
βββββChapter 7: AI Frameworks: Added headings and fixed image formatting in sections 7.1 and 7.2.βββββContributors: Updated content
π October 23
π Chapters
βββββChapter 7: AI Frameworks: Added links to frameworks when they are first introduced.βββββChapter 10: Model Optimizations: Added a section on efficient hardware implementation with corresponding images.βββββChapter 18: Robust AI: Added a placeholder for content related to Robust AI.βββββContributors: Updated content
π October 17
π Chapters
βββββChapter 7: AI Frameworks: Updated formatting for ml-frameworks sections.βββββChapter 18: Robust AI: Added a placeholder section for discussing the robustness of AI systems.βββββChapter 19: AI for Good: Added first draft of the AI for Good chapter content.βββββContributors: Updated content
π October 11
π Chapters
βββββChapter 3: DL Primer: Replaced callout-note with callout-tip for learning objectives.βββββChapter 5: AI Workflow: Replaced callout-note with callout-tip to enhance the visual clarity of learning objectives.βββββChapter 6: Data Engineering: Added learning objectives for the chapter.βββββChapter 7: AI Frameworks: Replaced callout-note with callout-tip for learning objectives.βββββChapter 8: AI Training: Changed callout style from βcallout-noteβ to βcallout-tipβ for learning objectives.βββββChapter 9: Efficient AI: Updated callouts for learning objects to be more informative.βββββChapter 10: Model Optimizations: Changed callout note style to callout tip for learning objectives.βββββChapter 11: AI Acceleration: Replaced callout-note with callout-tip for learning objects.βββββChapter 12: Benchmarking AI: Changed βcallout-noteβ to βcallout-tipβ for learning objectives.βββββChapter 13: ML Operations: Changed callout notes to callout tips for learning objectives.βββββChapter 14: On-Device Learning: Changed the type of callout used for learning objects from βcallout-noteβ to βcallout-tipβ.βββββChapter 15: Security & Privacy: Changed callout notes to callout tips for improved visual guidance of learning objectives.βββββChapter 16: Responsible AI: Learning object callouts were updated from βcallout-noteβ to βcallout-tipβ.βββββChapter 19: AI for Good: Learning objective callouts have been changed from βcallout-noteβ to βcallout-tipβ.βββββContributors: Updated contentβββββCase Studies: Replaced βcallout-noteβ with βcallout-tipβ for learning objectives.βββββEmbedded Ml: Updated callout notes to βcallout-tipβ for learning objectives.βββββEmbedded Sys: Changed callout notes to callout tips for learning objectives.βββββEthics: Updated callout style from βcallout-noteβ to βcallout-tipβ for learning objectives.βββββGenerative Ai: Changed callout note style to callout tip for learning objects.
π October 10
π Chapters
βββββChapter 6: Data Engineering: Added sections on data storage, version control, licensing, and a conclusion. Updated Data Processing and Data Sourcing sections based on feedback. Added a paragraph about Data Cascades and helpful references.βββββContributors: Updated contentβββββFront: Updated content
π October 08
π Chapters
βββββChapter 3: DL Primer: The chapter introduction was enhanced with learning objectives.βββββChapter 5: AI Workflow: Updated contentβββββChapter 9: Efficient AI: Fixed a broken reference.βββββChapter 11: AI Acceleration: Fixed a broken reference.βββββContributors: Updated contentβββββEmbedded Ml: Added learning objectives.βββββFront: Minor formatting adjustments were made to the navigation bar.βββββEmbedded Ml Exercise: Updated contentβββββEmbedded Sys Exercise: Updated content
π October 07
π Chapters
βββββChapter 1: Introduction: Minor text refinements were made.βββββChapter 3: DL Primer: Added placeholder for learning objectives.βββββChapter 5: AI Workflow: Added a placeholder for the learning objectives of this chapter.βββββChapter 6: Data Engineering: Added placeholder for learning objectives.βββββChapter 7: AI Frameworks: Added a placeholder section for learning objectives.βββββChapter 8: AI Training: Added placeholder for learning objectives.βββββChapter 9: Efficient AI: Added placeholder for learning objectivesβββββChapter 10: Model Optimizations: Added placeholder for learning objectivesβββββChapter 11: AI Acceleration: Added placeholder for learning objectivesβββββChapter 12: Benchmarking AI: Added placeholder for learning objectivesβββββChapter 13: ML Operations: Added placeholder for learning objectives.βββββChapter 14: On-Device Learning: Added placeholder for learning objectives.βββββChapter 15: Security & Privacy: Added placeholder for learning objectives.βββββChapter 16: Responsible AI: Added a placeholder for learning objectives.βββββChapter 19: AI for Good: Added placeholder for learning objectives.βββββContributors: Updated contentβββββEmbedded Ml: Added exercises to reinforce learning concepts and included placeholders for specified learning objectives.βββββEmbedded Sys: Added exercises based on feedback and included placeholders for learning objectives.βββββEmbedded Ml Exercise: Updated contentβββββEmbedded Sys Exercise: Updated contentβββββTest: Updated contentβββββ** Embedded Ml Exercise**: Updated contentβββββCase Studies: Added placeholder for learning objectivesβββββEthics: Added placeholder for learning objectives.βββββGenerative Ai: Added a placeholder for learning objectives.
π September 30
π Chapters
βββββChapter 9: Efficient AI: Updated contentβββββChapter 10: Model Optimizations: Added section headers for improved readability.βββββChapter 11: AI Acceleration: Added section headers for improved readability.βββββContributors: Updated contributor list.
π September 29
π Chapters
βββββChapter 9: Efficient AI: Added a draft overview section for the efficient AI chapter.βββββChapter 11: AI Acceleration: Added an initial draft of the AI acceleration section with a focus on providing an overview of the topic.βββββChapter 17: Sustainable AI: Updated contentβββββContributors: Updated content
π September 28
π Chapters
βββββChapter 1: Introduction: Added a section on AI for social good with examples in healthcare and education.βββββAi Social Good: Outlined the structure for an AI for social good section.βββββContributors: Updated the list of contributors.βββββIndex: Added section on AI for social good with examples of applications in healthcare, education, and environmental sustainability.
π September 27
π Chapters
βββββChapter 1: Introduction: Added a section discussing how AI can be used for social good.βββββChapter 7: AI Frameworks: Updated the frameworks section outline.βββββChapter 11: AI Acceleration: Improved chapter organization by folding a skeleton section on emerging hardware into the existing AI acceleration chapter.βββββAi Social Good: Added an outline for the AI for social good section.βββββContributors: Updated contentβββββIndex: Added AI for social good section with examples of applications in healthcare and education.
π September 24
π Chapters
βββββChapter 3: DL Primer: Resolved instances where references were unintentionally removed during copyediting.βββββChapter 12: Benchmarking AI: Placeholder for talking about data benchmarkingβββββContributors: Updated contentβββββEmbedded Sys: Updated content