Trustworthy LLMs: Principles and Challenges

Mohit Kumar

Mar 2027 · Morgan Kaufmann

Ebook

250

Pages

Eligible

This book will become available on March 1, 2027. You will not be charged until it is released.

About this ebook

Despite numerous advantages, LLMs have trust, transparency, accountability, and reliability issues due to development with "black-box" approaches, which make it difficult to understand how LLMs create specific outputs. Trustworthy LLMs: Principles and Challenges presents the fundamental concepts of trustworthy LLMs, then proceeds to address the foremost challenges researchers and developers face in developing reliable and trustworthy LLMs. The book begins by presenting the main research branches of artificial intelligence along with the principles of LLMs, from pre-training to fine tuning, and, ultimately, trustworthy LLMs. Readers will learn about the chief technical principles of LLMs, including attention mechanism, transformers, and transfer learning. The methodologies used for development of ChatGPT have been explained as a case study for comprehensive understanding of the concepts involved in LLMs. Readers will also learn about the integration of XAI with LLM, and other key frontiers in trustworthy LLM development, including the synergy between deep learning and LLMs, as well as case studies on GPT-4 and OPT-1.3B. The book concludes with chapters on key challenges and future research approaches for developing trustworthy LLMs. - Provides a well-organized framework to understand how Large Language Models work, and how they are evolving - Offers a structured mapping of LLM methodologies, including transformers, attention mechanism, and transfer learning - Presents the latest technical frameworks combined with real-world examples and case studies such as ChatGPT, TrustGPT, GPT-4, and OPT-1.3B - Provides a comprehensive reference on trustworthy LLMs, along with the challenges and future research frontiers for their development

Discover more

About the author

Mohit Kumar, PhD is Assistant Professor in the Department of Information Technology at Dr. B R Ambedkar National Institute of Technology, Jalandhar, India. He received his Ph.D. degree from Indian Institute of Technology Roorkee in the field of Artificial Intelligence and Cloud Computing, 2018, and M. Tech degree in Computer Science and Engineering from ABV-Indian Institute of Information Technology Gwalior, India in 2013. He has received his B. Tech degree in Computer Science and Engineering from MJP Rohilkhand University Bareilly, 2009. His research topics cover the areas of Cloud computing, Fog/ Edge Computing, Internet of Things, federated learning, Blockchain, and Artificial Intelligence. Dr Mohit received best faculty award in NIT Jalandhar for academic session 2022-2023. He has published more than 100 research articles in reputed journals, IEEE Transactions and international conferences. He has been Session chair and keynotes Speaker of many International conferences, webinars, FDP, STC in India. He has guided six M. Tech Thesis and guiding 6 Ph.D. Scholars. He has been listed in the prestigious Top 2% of Scientists in the world (2023, 2024) announced by Elsevier and Stanford University, United States. He is editorial board member of several reputed journals such as scientific report (SCIE), discover computing (SCIE) and Discover Artificial Intelligence (Scopus indexed). He is an active reviewer of several reputed journals and international conferences. He is a member of the IEEE.

Reading information

Smartphones and tablets

Install the Google Play Books app for Android and iPad/iPhone. It syncs automatically with your account and allows you to read online or offline wherever you are.

Laptops and computers

You can listen to audiobooks purchased on Google Play using your computer's web browser.

eReaders and other devices

To read on e-ink devices like Kobo eReaders, you'll need to download a file and transfer it to your device. Follow the detailed Help Center instructions to transfer the files to supported eReaders.

Flag as inappropriate