New
Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow
A practical, project-driven introduction to machine learning and deep learning with Python
2022
View →
Pages
1167
Published
2021
AI Learning
✨ New
Artificial Intelligence: A Modern Approach, Global Edition
The definitive textbook on intelligent systems, from foundational search and logic to modern machine learning and probabilistic reasoning
Build a rigorous, end-to-end understanding of AI — from search algorithms and knowledge representation to reinforcement learning and the ethics of intelligent systems.
Russell and Norvig's Artificial Intelligence: A Modern Approach is the standard reference for the field, used in university courses and research labs worldwide. Spanning 1,167 pages across every major subdiscipline of AI, it gives you the theoretical foundations and algorithmic tools to understand, evaluate, and build intelligent systems. Whether you are approaching AI for the first time or shoring up gaps in your working knowledge, this is the book the field measures itself against.
About this book
There is one book that every serious AI practitioner has on their shelf, or wishes they did. Russell and Norvig's Artificial Intelligence: A Modern Approach has defined the curriculum for AI education since its first edition, and this Global Edition brings the content fully up to date for the modern landscape of machine learning, deep neural networks, probabilistic models, and autonomous agents.
The book does not chase trends. It builds understanding from first principles: what it means for an agent to act rationally, how search and planning algorithms find solutions in complex state spaces, how logic and probabilistic reasoning let a system model the world, and how learning from data connects statistical theory to practical inference. Each chapter is self-contained enough to read on its own, yet the whole forms a coherent map of the field.
Coverage spans a remarkable breadth without sacrificing depth. You will work through classical and heuristic search, constraint satisfaction, propositional and first-order logic, Bayesian networks, hidden Markov models, decision-theoretic planning, reinforcement learning, supervised and unsupervised learning, natural language processing, computer vision, robotics, and multi-agent systems. The final sections address the philosophical and ethical dimensions of AI — questions that have become impossible to ignore as these systems move into production.
- Rigorous algorithmic treatments with pseudocode you can implement directly
- Worked examples and exercises at the end of every chapter
- Unified agent-based framework that ties every topic to rational decision-making
- Updated coverage of deep learning, large-scale probabilistic inference, and modern NLP
- Attention to real-world applications alongside theoretical guarantees
This is not a tutorial or a project guide. It is a reference-grade textbook. Read it cover to cover as a structured course, or keep it at your desk and open it when you need to understand why a technique works, not just how to call it from a library. Either way, it earns its place on the shelf.
🎯 What you'll learn
- Formalize any problem as a search or planning task and choose the right algorithm for its structure
- Build and query Bayesian networks to reason correctly under uncertainty
- Apply supervised, unsupervised, and reinforcement learning techniques with a grounded understanding of their statistical basis
- Represent knowledge in propositional and first-order logic and use automated inference to draw conclusions
- Design rational agents that perceive, plan, and act across fully and partially observable environments
- Evaluate the tradeoffs between exact and approximate inference in probabilistic models
- Reason about the ethical, philosophical, and safety implications of deploying intelligent systems
👤 Who is this book for?
- Computer science students taking an AI course who want the primary text the syllabus is built around
- Software engineers moving into machine learning or AI roles who need rigorous foundations, not just framework tutorials
- Data scientists who use probabilistic models daily but want to understand the theory connecting Bayesian networks, HMMs, and decision processes
- Researchers entering AI-adjacent fields who need a reliable map of what the discipline covers and where its open problems lie
- Practitioners preparing for graduate study or technical interviews where algorithmic AI knowledge is tested in depth
Table of contents
-
01
Introduction and the Foundations of AI
Defines what artificial intelligence is, surveys its history, and establishes the rational agent framework that unifies every topic in the book. You will understand why rationality, not human imitation, is the right design target.
-
02
Intelligent Agents
Introduces the agent abstraction: percepts, actions, environments, and performance measures. You will classify environment types and match them to appropriate agent architectures before writing a line of algorithm.
-
03
Search in Deterministic Environments
Covers uninformed and informed search algorithms including BFS, DFS, uniform-cost search, A*, and greedy best-first search. You will analyze their completeness, optimality, and time and space complexity on concrete problems.
-
04
Search in Complex Environments
Extends search to local, online, and adversarial settings including hill climbing, simulated annealing, genetic algorithms, and minimax with alpha-beta pruning. You will apply these to games and optimization problems with incomplete information.
-
05
Constraint Satisfaction Problems
Formalizes CSPs and presents backtracking search, arc consistency, and local search methods for solving them. You will solve scheduling and configuration problems by choosing and combining the right propagation and search strategies.
-
06
Logical Agents and Knowledge Representation
Introduces propositional logic, model checking, and resolution-based inference. You will build a knowledge base that lets an agent derive new facts from stored beliefs using sound, complete reasoning procedures.
-
07
First-Order Logic and Automated Reasoning
Lifts propositional logic to predicates, quantifiers, and unification, then covers forward chaining, backward chaining, and resolution. You will represent complex domains and run automated inference over them.
-
08
Probabilistic Reasoning and Bayesian Networks
Develops the probability theory underlying uncertainty in AI, then builds Bayesian networks and covers exact and approximate inference algorithms. You will model real-world uncertainty and compute posterior distributions over hidden variables.
-
09
Learning from Data
Covers supervised learning from decision trees and linear models through neural networks and deep learning, plus unsupervised clustering and dimensionality reduction. You will connect statistical learning theory to the algorithms used in practice.
-
10
Reinforcement Learning, Applications, and the Future of AI
Presents temporal-difference learning, Q-learning, and policy gradient methods, then surveys robotics, NLP, vision, and multi-agent systems before addressing the philosophical and ethical stakes of advanced AI. You will finish with a coherent picture of where the field stands and where it is heading.
Frequently asked questions
Do I need a strong math background to read this book?
Comfort with calculus, linear algebra, and basic probability will let you get full value from the technical chapters. Many sections can be read for conceptual understanding without working through every derivation, but the more math you bring, the more you take away.
Is this a project-based book with code to download?
No. This is a rigorous academic textbook, not a hands-on tutorial. Algorithms are presented in pseudocode designed to be readable and implementable, but there is no companion code repository. Readers looking for guided projects should pair it with a practical supplement.
How does this Global Edition differ from the US edition?
The Global Edition contains the same core content as the fourth US edition, published by Pearson in 2021, with minor formatting and packaging differences. The technical content, chapter structure, and exercises are equivalent.
Is this suitable for self-study, or is it designed for a classroom?
It works well for self-study. Each chapter includes exercises with varying difficulty, and the agent-based framework gives the book a coherent thread you can follow independently. Many readers work through selected chapters based on their own gaps rather than reading sequentially.
Does this edition cover deep learning and modern neural networks?
Yes. The fourth edition substantially expanded coverage of deep learning, large-scale probabilistic inference, and contemporary NLP compared to earlier editions, reflecting the field's shift since 2010.
Who should not buy this book?
If you are looking for a quick practical guide to using a specific framework like PyTorch or TensorFlow, this is not that book. It provides the theoretical foundations that make frameworks intelligible, not step-by-step project walkthroughs.
You might also like
New
New
Designing Machine Learning Systems
An Iterative Process for Production-Ready Machine Learning Applications
by Chip Huyen
2022
View →
New
Probabilistic Machine Learning
A rigorous foundation in Bayesian reasoning, probabilistic models, and modern machine learning methods
2022
View →
New
AI and Machine Learning for Coders
A Programmer's Guide to Building AI and Machine Learning Models with TensorFlow
2020
View →