Customer Reviews

Artificial Intelligence: A Modern Approach

5 star:		(46)
4 star:		(22)
3 star:		(11)
2 star:		(7)
1 star:		(2)

 

 

I didn't think that the first edition of this book was as bad as some of the reviewers said, but the second edition is definitely a vast improvement. It's not just some obligatory 2nd edition that some authors release to say that they are staying actively published. The first edition was somewhat confusing in its explanations and the exercises were really blurry on what was being asked. All of that has now been resolved.
The book is a comprehensive and insightful introduction to artificial intelligence with an academic tone. It provides a unified view of the field organized around the rational decision making paradigm, which focuses on the selection of the "best" solution to a problem. The book's overall theme is that the purpose of AI is to solve problems via intelligent agents, and then goes about specifying the features such an agent or agents should have. Pseudocode is provided for all of the major AI algorithms. Being about the broadest book in terms of coverage of AI, you should therefore not expect it to be the deepest in coverage. However, each topic is covered to the extent that the reader should understand its essence. Sections one through six are absolutely wonderful, and comprise the "meat" of AI. Section seven is rather weak since it tries to cover both robotics and text processing in their own individual chapters, and entire books have a hard time covering this material. Section eight is different from the others, since it talks about the philosophy and future of AI.
Another plus for this book is that there is a great deal of extra material that deals with standard AI curriculum. For example, the chapters on logic not only include the typical introduction to propositional and first order logic together with the usual inference procedures, they also give many useful hints how to use first order logic to actually represent aspects of the real world such as measures, time, actions, mental objects, etc. These chapters also contain much information about how to implement efficient logical reasoners.
Finally, this second edition has an excellent website that can be found by going through the publisher's webpage for the book. This website contains four sample chapters, pseudocode, and actual code in Java, Python, and LISP.
I notice that Amazon shows the table of contents from the first edition, so I am showing what the actual table of contents is for the second edition for the purpose of completeness. Note that the book has been significantly reorganized.
I. ARTIFICIAL INTELLIGENCE.
1. Introduction.
2. Intelligent Agents.
II. PROBLEM-SOLVING.
3. Solving Problems by Searching.
4. Informed Search and Exploration.
5. Constraint Satisfaction Problems.
6. Adversarial Search.
III. KNOWLEDGE AND REASONING.
7. Logical Agents.
8. First-Order Logic.
9. Inference in First-Order Logic.
10. Knowledge Representation.
IV. PLANNING.
11. Planning.
12. Planning and Acting in the Real World.
V. UNCERTAIN KNOWLEDGE AND REASONING.
13. Uncertainty.
14. Probabilistic Reasoning Systems.
15. Probabilistic Reasoning Over Time.
16. Making Simple Decisions.
17. Making Complex Decisions.
VI. LEARNING.
18. Learning from Observations.
19. Knowledge in Learning.
20. Statistical Learning Methods.
21. Reinforcement Learning.
VII. COMMUNICATING, PERCEIVING, AND ACTING.
22. Agents that Communicate.
23. Text Processing in the Large.
24. Perception.
25. Robotics.
VIII. CONCLUSIONS.
26. Philosophical Foundations.
27. AI: Present and Future.

Thanks to all those who reviewed the first edition.
If you are reading this, you will probably want the
second edition instead. It was published Dec 20, 2002.
Every chapter has been extensively rewritten.
Significant new material has been introduced to cover
areas such as constraint satisfaction, fast propositional
inference, planning graphs, internet agents, exact
probabilistic inference, Markov Chain Monte Carlo
techniques, Kalman filters, ensemble learning methods,
statistical learning, probabilistic natural language
models, probabilistic robotics, and ethical aspects of AI.

For more information, see aima.cs.berkeley.edu

Most of the reviews here refer to the first edition, not the second. There have been significant changes to the second edition. Amazon should consider using a display that clarifies which edition each review refers to.

Russell and Norvig's AI: A Modern Approach is THE best AI text out there. At 932 pages it is encyclopedic, it has nearly everything. So what is missing? How could it be improved? Probably the worst thing about the book is the binding. I am not sure that you can read it from cover to cover without some pages coming loose. Perhaps its the length. Perhaps it needs to be split into two volumes. I am not a fan of pseudocode and all the algorithms are in pseudocode. I think the right compromise between detailed practical code and tutorial compactness is something like the code in Jackson's text Expert Systems. I realize this might make a long book even longer but I still think some examples in Lisp, Prolog, etc. would be an improvement. There are a few things missing. Some detail on case-based reasoning is needed and some newer topics like hybrid systems and rough sets. Also, more on parallel computing for AI. Occasionally I was annoyed by the references. On page 27 the authors attribute a story to Heckerman's 1991 thesis. The thesis contains no such story. The reference should have been to a private communication. By now you might think I hate the book. No. I am suggesting improvements. I repeat. It is THE BEST SINGLE AI TEXT IN PRINT. But you will not be able to teach the whole book in a single AI course. Not even a two semester course.

Until recently, my Algorithms book was my favorite text book ever. However, AI: A Modern Approach has supplanted it. This book is the most thoughtfully designed, easily understandable, clear text I've ever used in over 28 years of attending schools. I really knew nothing about AI when I took my first grad class in AI, but this book, along with a pretty great instructor, has been a wonderful resource, more than any other book I've used. I have not need to google for more information or speak to the professor. The answers are here--clear and concrete.

Have no fear and trust this book!

While R&N's coverage of topics in Artificial Intelligence is no doubt encyclopedic, several problems exist with the book:

1) The textbook is awfully traditional and only mentions in passing newer trends in AI. For example, case-based reasoning (or the "Yale view of AI") is mentioned, but not covered. Because AI is a new and rapidly changing field, and because AI paradigms are usually based on a small set of ontological assumptions, I believe it would not be too difficult for students to understand new paradigms. Obviously this should be a high pedagogical priority.

2) The textbook is rather condescending, with the authors strongly imposing their viewpoints. In other words, the authors are a little too dogmatic and that is reflected in the text. For example, they sometimes go about ranking paradigms.

3) The textbook is sometimes rather ambiguous when explicating certain paradigms, and the end-of-chapter problems are very, very ambiguous. One of the justifications for unclear questions is to get people thinking, but when the theoretical explications are already ambiguous it defeats the whole purpose.

4) The philosophical sections in R&N are rather naive and superficial.

In spite of its obvious shortcomings, R&N has been tremendously useful to me, and I recommend it as a reference. The good news, I've heard, is that a new edition of R&N is coming out next year where these problems are eliminated.

The book offers mathematically based coverage of the most important topics in AI. It has a very pragmatic view and does not spend too much time with the over-hyped topics, such as fuzzy logic, genetic algorithms. Instead, it provides clear and easy to understand examples for most of the introduced concepts.

The text requires some knowledge of computer science, mathematics, and statistics. Sections on topics that require extensive outside knowledge, such as computational learning theory, are not covered very deeply, but usually offer nice insight to the concepts and basic principles.

I recommend this book both for beginners to exact AI as an introductory text and to more experienced researchers as an invaluable reference. I found it to be the most useful book I own.

Progress in the field of artificial intelligence has executed a random walk after establishing itself with a bang in the 1950s. Optimistic predictions of the future of A.I. in that decade only partially came true in the decades after that. Currently, the field is divided up into subfields going by the names data mining, computational intelligence, intelligent agent theory, expert systems, etc. This book is the best book available for learning about this fascinating and important subject. The applications of A.I. are enormous, and will increase dramatically in the decades ahead. Indeed the prospects are very exciting, and the authors themselves have been involved heavily in extending the frontiers of the subject. Some of the main points of the book that really stand out include:
1. The useful exercises at the end of each chapter. 2. The discussion of simple reflex and goal-based agents. 3. The treatment of constraint satisfaction problems and heuristics for these kinds of problems. 4. The overview of iterative improvement algorithms, particularly the discussion of simulated annealing. 5. The discussion of propositional logic and its limitations as an effective A.I. paradigm. 6. The treatment of first-order logic and its use in modeling simple reflex agents, change, and its use in situation calculus. There is a good overview of inference in first-order logic in chapter 9 of the book, including completeness and resolution. 7. The treatment of logic programming systems; the Prolog language is discussed as a logical programming language. Noting that Prolog cannot specify constraints on values, the authors discuss constraint logic programming (CLP) as an alternative logic programming language that allows constraints. 8. The discussion of semantic networks and description logics. 9. The treatment of conditional programming via the conditional partial-order planner (CPOP). 10. Representing knowledge in an uncertain domain and the semantics and inference in belief networks. 11. The brief discussions on stochastic simulation methods and fuzzy logic. 12. The discussion on computational learning theory 13. The treatment of neural networks, especially the discussion of multilayer feed-forward networks and the comparison between belief networks and neural networks. 14. The brief discussion on genetic algorithms and evolutionary programming. 15. The discussion on explanation-based learning and the technique of memoization. 16. The (excellent) overview of inductive logic programming. This relatively recent area was new to me at the time of reading so I appreciated the discussion. The authors briefly mention the approach of discovery systems and the Automated Mathematician (AM). 17. The interesting discussion of telepathic communication between robots via the exchange of internal representations. 18. The discussion on a formal grammar for a subset of English and the extensive treatment of natural language processing. 19. The discussion of speech recognition and the use of hidden Markov models and the Viterbi algorithm. 20. The fascinating discussion on robotics, particularly the treatment of configuration spaces, which brings in some techniques from computational geometry and topology. 21. The discussion on the philosophical ramifications of A.I. Future developments in A.I. will provide a unique testing ground for philosophy, in a way that will be unparalleled in the history of philosophy. Philosophers critical of A.I. will have the opportunity to check whether their arguments against the possibility of "strong A.I.", are in fact true.

The publication of this textbook was a major step forward, not only for the teaching of AI, but for the unified view of the field that this book introduces. Even for experts in the field, there are important insights in almost every chapter. I recommend it to anyone who wants to have an introductory overview of the state of AI. And I recommend it to experts in the field, who will enjoy its unified description of the field. I especially enjoyed the introductory chapter and the chapter on philosophical issues. I have taught from this book three times now, and it has improved my AI class hugely.

Have surveyed various AI books for lecturing of AI (Foundation) and found that this is the best book as far as the syllabus is concerned: Introduction, Intelligent Agents (v good), Search, Knowledge Representation, Machine Learning, Planning, NLP and Philosophy of AI. The rich set of examples are also useful and the students enjoyed them. Having said that, specific areas such as NLP, Neural Networks, Machine Vision are better off in specialised book. Would be great if there are references to specific websites for more further details.