NNDL is a textbook series on neural networks and deep learning available in two versions designed for different audiences and learning goals.
The project addresses the need for accessible yet rigorous education in deep learning theory. It offers two parallel reading paths: a theoretical version that systematically covers machine learning fundamentals, classical neural networks, optimization, Transformers, graph neural networks, reinforcement learning, generative models, large language models, and intelligent agents; and a general knowledge version that prioritizes intuition over formalism, using stories, case studies, and analogies to explain the same concepts while extending into multimodal learning, scientific AI, embodied intelligence, and AI safety. Readers can choose based on their background and goals, or combine both approaches.
The theoretical version suits professional courses, research entry points, and those seeking a complete theoretical framework. The general knowledge version targets non-specialists, interdisciplinary learners, university general education courses, and those wanting rapid comprehension of modern AI. The project explicitly recommends pairing either version with a companion practice-focused repository for hands-on coding experience. Both versions are currently in pre-publication stages as electronic drafts undergoing continuous updates, with final print editions pending publisher announcements.
Issues and pull requests often wait weeks or longer for a first response.
