Explainable AI
Description
With the increasing use of deep learning systems across various industries, there is a growing need to make their decision-making processes more understandable and transparent. Regulatory requirements now demand clarity, and users and stakeholders want to know how AI systems work. The textbook addresses these needs by providing a detailed guide on integrating Explainable AI (XAI) into the Deep Learning Operations (DLOps) pipeline. By doing so, organizations can implement Continuous Integration (CI) and Continuous Deployment (CD) practices effectively.
Explainable AI: Building Trustworthy Deep Learning Systems focuses on how to incorporate XAI models, tools, and techniques to clarify machine learning decisions. It explores applications in fields such as healthcare, defense, human activity recognition, and object identification. The book offers practical advice on embedding XAI tools throughout the lifecycle of deep learning systems, covering topics like Explainability and Interpretability, Deep Learning Operations (DLOps), and Machine Learning Operations (MLOps). It also includes real-world examples, challenges, and solutions.
This textbook is ideal for undergraduate and graduate students studying computer science, electronic and communications engineering, and electrical and electronics engineering. It is particularly suited for courses like AI Internals in Cyber-Physical Systems, AI Security Analytics, and Human-Computer Interaction with XAI. Professionals in systems engineering and industrial engineering will also find it valuable.
For those adopting the textbook for courses, a solutions manual and PowerPoint slides are available.
More details
Other editions
Additional editions
Persons
Dr. D. Jeya Mala, a professor at the School of Computer Science and Engineering (SCOPE), Vellore Institute of Technology, Chennai, India, has over 21 years of teaching and research experience and four years in industry. She is an expert evaluator for AICTE-NEAT Cell and a member of the National Work Group on Quantum Computing formed by TCCI, Govt. of India. Holding a design patent and a published utility patent, she has authored over 65 papers in SCI and Scopus-indexed journals, conferences, and book chapters. Recognized globally, she is listed in the SEBASE repository of the University College of London for her work in Search-Based Software Engineering and has received awards such as the "Best Faculty of the Year 2024" from the Computer Society of India, the "Certificate of Excellence" and "Top Performer Award" from IIT Bombay, and the "Faculty Research Excellence Award" for three consecutive years. She serves on review boards, editorial boards, and technical committees for reputed journals and conferences and is affiliated with IEEE, ACM, the Indian Science Congress Association, the Computer Society of India, i-Soft, and Machine Intelligence Research Labs. Her research focuses on Quantum Computing, Artificial Intelligence, Explainable AI, Machine Learning, Deep Learning, and Analytics.
Dr. Subramaniam Ganesan, a Professor in the Department of Electrical & Computer Engineering (ECE) at Oakland University, Rochester, MI, USA, is a distinguished academic and researcher with extensive contributions to his field. A senior member of IEEE, he has served as an IEEE Computer Society Distinguished Visiting Speaker, an IEEE Region 4 technical activities member, and is a Fellow of ISPE. Dr. Ganesan has been recognized with numerous prestigious awards, including the Lifetime Achievement Award from ISAM, the Lloyd L. Withrow Distinguished Speaker Award from SAE, the Best Paper Award from ISAM, and teaching excellence awards from ASEE and Oakland University. As the editor-in-chief of two internationally renowned journals, he has also organized the "Systems Engineering" panel at the SAE World Congress for 15 consecutive years. With several patents in embedded systems, his research interests span Real-Time Systems, Parallel Architecture, Mobile Computing, Automotive Embedded Systems, and Signal Processing.
Content
Section I: Foundations on XAI. 1. Introduction to XAI Taxonomy, Ethics and Policy. 2. XAI Techniques and Models. 3. Intrinsically Interpretable Models. 4. Posthoc Explainability Models. Section II: Integration of XAI into DLOps. 5. DLOps - Introduction to XAI in DL Model Development. 6. Integrating XAI into Model Training. 7. XAI Integration in Model Validation. 8. XAI integration in CI/CD. 9. XAI Integration in Model Deployment. Section III: Emerging Trends in XAI. 10. Integrating XAI in Agentic AI Architecture. 11. Interplay of XAI with LLM.