Natural Language Processing: Neural Networks and Large Language Models - Tong Xiao

This book offers a comprehensive and forward-looking introduction to Natural Language Processing (NLP). Moving beyond traditional NLP approaches, this book presents a unified framework that connects foundational methods with the breakthroughs enabled by Large Language Models (LLMs).

Designed for students, engineers, and researchers in computer science and artificial intelligence, this volume bridges the gap between classic NLP concepts and modern deep learning paradigms. Part I introduces the core principles of machine learning and neural networks, building the foundation for understanding representation learning. Part II examines key neural architectures—word vectors, recurrent and convolutional models, sequence-to-sequence systems, and the Transformer—that have revolutionized NLP in the past decade. Part III brings readers to the cutting edge, covering pre-training, generative modeling, prompt design, alignment, and inference in LLMs.

By blending theory with practical insight, this book helps readers grasp both how and why neural NLP works. It answers essential questions: What makes LLMs fundamentally different from earlier models? How do they generate, align, and reason about text? Whether used as a textbook or as a technical reference, it provides a structured path to mastering natural language processing in the deep learning era.