Images in this review
Reviews with images
The 4th edition is a welcomed update to the definite AI textbook.
The 4th edition of Artificial Intelligence: A Modern Approach (AIMA) is a welcomed update to the definite AI textbook. The timing is right, given the renewed interest and recent rapid advances in Artificial Intelligence (AI). AIMA strives to cover the breadth and depth of AI with a focus on practical applications. The book gives fair coverage to the diverse field of AI in its 1,000+ pages. This edition has an increased emphasis on machine learning, uncertain and evolving objectives, and ethics. It is a pleasant read. The writing is easy to follow. The writers build up complex AI concepts piece-by-piece. The examples are contemporary and applied. Highlights include stellar contributing guest writers. For example, Ian Goodfellow contributed to the chapter on deep learning and Anca Dragan contributed to the robotics chapter.
Thank you for your feedback Sorry, there was an error
Sorry we couldn't load the review
Reviewed in the United States on January 13, 2021 - I liked the structure of the book in terms of content and printing format.
7 people found this helpful Reviewed in the United States on August 4, 2022 Very clearly written with all arguments straightforwardly expressed. I found it most helpful to cover in detail what we mean by intelligence/thought/capability before diving into the mechanics. Also the authors are honest about what is not known, including why certain methods work better than others.
6 people found this helpful Reviewed in the United States on February 12, 2021 Has everything a beginner needs to know and master AI. Also easy to read and understand. Finally the author has created a website for this book where implementation of real life AI problems are written in java, python and various other languages.
3 people found this helpful Reviewed in the United States on March 12, 2021 George Polya, the mathematician, once wrote: "If you can't solve a problem, then there is an easier problem you can solve. Find it". This book helps you find the easier problem. This is the greatest complement I could give to any technical book.
8 people found this helpful Reviewed in the United States on November 20, 2020 I like that this book covers a lot of ground and provides super useful information and algorithms. Super great book. I don't like that some parts in the algorithm pseudo code aren't thoroughly explained with examples of usage. Also, the formulae notation for the i'th and j'th iteration characters aren't properly explained or consistent across formulae, making it more difficult for beginners. I think the formulae notations could be more consistent and thoroughly explained.
10 people found this helpful Reviewed in the United States on June 28, 2024 Specialized fields attract positive reviewer bias for the same reason that interlocking boards attract ethical scrutiny: they are comprised of near peers who may need each other in the future. Even I, not a remote pretender to the poor branch of the family, feel this tug, that somehow, somewhere, I will enter into the sphere of influence of Russell & Norvig and my lukewarm review will come back to haunt me. I will have to take that chance. The intended captive audience is comprised of (or is it "comprises"?) students who must use this book and be tested on material in it, and it shows. Ideal academic material justifies its existence by providing difficult, testable subject matter, perhaps on the theory that exercise is exercise, much as prospective colonial administrators during the age of British empire were subjected to the rigor of ancient Latin and Greek—well understood, difficult, and testable—so those so winnowed would likely be up to running colonies. Maybe it was even an effective system. The equivalent observation here stems from material like this: "The LRTA∗ algorithm was developed by Korf (1990) as part of an investigation into realtime search for environments in which the agent must act after searching for only a fixed amount of time (a common situation in two-player games). LRTA∗ is in fact a special case of reinforcement learning algorithms for stochastic environments (Barto et al., 1995)." "Common in two-player games"? As opposed to what!? Except for the sometimes open-ended limit of how long to look in a particular case it is the universal situation in human experience that we must act under imperfect information after a finite period of evaluation; this was understood before Korf devised a special algorithm to handle a specialized case. Then there is: "stochastic environments". This is academic speak for "environments which contain an element of chance". "Stochastic" dresses this universal situation up so it sounds like we said something sophisticated. We didn't, we just displayed a few ostentatious feathers. Well, I did get a textbook, whereas the more efficient method of investigation is interrogation. What did I expect?
3 people found this helpful Reviewed in Japan on June 12, 2021 Some figures are missing, and mathematical expressions are messed up. Reviewed in Germany on February 8, 2021 Meine Tochter studiert artifical intelligence online an der IUBH. sie ist von dem Buch begeistert. Es ist sehr unformativ und für ihr Studium sehr hilfreich. Reviewed in Spain on March 22, 2021 Este libro ha sufrido varias reediciones, pero la edición 2020 es simplemente soberbia, es un tratado completo de todas las áreas de la Inteligencia Artificial, completamente actualizada. 5.0 out of 5 stars
Reviewed in Spain on March 22, 2021
Este libro ha sufrido varias reediciones, pero la edición 2020 es simplemente soberbia, es un tratado completo de todas las áreas de la Inteligencia Artificial, completamente actualizada.
Reviewed in the United Kingdom on May 22, 2021 For anyone who has studied the previous version(s) of this book, you'll know just how detailed and incredibly comprehensive this is. It is generally agreed to be the most credible and thorough book within the field of AI, by virtue of the backgrounds of the primary authors, along with its vast size of 1100+ pages. Even at this vast length, the information distilled into each of the sections is dense and very valuable, and I cant really highlight any chapters that are particularly redundant. You can also gain an incredible amount of knowledge from each of the sections of the book, even without a full understanding of the concepts the first time round (or reading many of the earlier chapters). For this reason it is widely used as a primary reference throughout academic courses in AI. Where this book really shines is its heavy focus on foundational AI principles and topics that are concrete and timeless (as timeless as concepts can be in such a fast-moving field!). The modernised field of AI is heavily dominated by machine learning, but this represents only a subset of the field, with a vast expanse of other important subfields. This book branches across into all of these other areas (along with strong coverage of machine learning too), and if studied, will provide you with a very strong and grounded foundation in AI. It should be highlighted that the book is challenging, and is far from a simple read. It is very good as a reference book, and for dipping into and out of as required - it's unlikely you'll manage to commit to reading the book from start to finish. This is not due to fault of the authors, who do a fantastic job of using engaging and easily-read writing styles, but is simply by virtue of the complicated and vast topics that the book is based on. In terms of the 4th Edition itself, I would fully recommend this updated version over any of the previous versions. There is a significant amount of new material and improvements compared to the 3rd edition, which helps capture the major developments throughout the past ten years. There are now extensive chapters on Deep learning, probabilistic programming, multi-agent architectures, natural language processing, computer vision and robotics. Furthermore, the book now has a much better glossary with huge range of topics to quickly find, which was definitely a downside of prior editions. In terms of the physical book itself, I have no major issues with the 4th Edition and can summarise as follows: Overall, I thoroughly recommend. An essential book for the collection of any AI researchers, students, data scientists or AI practitioners. 5.0 out of 5 stars
Reviewed in the United Kingdom on May 22, 2021
For anyone who has studied the previous version(s) of this book, you'll know just how detailed and incredibly comprehensive this is. It is generally agreed to be the most credible and thorough book within the field of AI, by virtue of the backgrounds of the primary authors, along with its vast size of 1100+ pages. Even at this vast length, the information distilled into each of the sections is dense and very valuable, and I cant really highlight any chapters that are particularly redundant. You can also gain an incredible amount of knowledge from each of the sections of the book, even without a full understanding of the concepts the first time round (or reading many of the earlier chapters). For this reason it is widely used as a primary reference throughout academic courses in AI. Where this book really shines is its heavy focus on foundational AI principles and topics that are concrete and timeless (as timeless as concepts can be in such a fast-moving field!). The modernised field of AI is heavily dominated by machine learning, but this represents only a subset of the field, with a vast expanse of other important subfields. This book branches across into all of these other areas (along with strong coverage of machine learning too), and if studied, will provide you with a very strong and grounded foundation in AI. It should be highlighted that the book is challenging, and is far from a simple read. It is very good as a reference book, and for dipping into and out of as required - it's unlikely you'll manage to commit to reading the book from start to finish. This is not due to fault of the authors, who do a fantastic job of using engaging and easily-read writing styles, but is simply by virtue of the complicated and vast topics that the book is based on. In terms of the 4th Edition itself, I would fully recommend this updated version over any of the previous versions. There is a significant amount of new material and improvements compared to the 3rd edition, which helps capture the major developments throughout the past ten years. There are now extensive chapters on Deep learning, probabilistic programming, multi-agent architectures, natural language processing, computer vision and robotics. Furthermore, the book now has a much better glossary with huge range of topics to quickly find, which was definitely a downside of prior editions. In terms of the physical book itself, I have no major issues with the 4th Edition and can summarise as follows: Overall, I thoroughly recommend. An essential book for the collection of any AI researchers, students, data scientists or AI practitioners. Reviewed in Saudi Arabia on November 27, 2025 طلبت نسخه اصليه غلاف صلب ارسلتو ورق عادي غلاف عادي
Top reviews from the United States
There was a problem filtering reviews. Please reload the page.
5.0 out of 5 stars
Very succinct book
- I did not like the binding; although it is hardcover (which I prefer over paperback), it is glued like a paperback, and I am concerned the pages may start coming apart.
- I would have liked some examples within the text, as well as exercises (along with solutions in a separate instructor's manual).
5.0 out of 5 stars
Exceptionally well written - balancing the theoretical and practical
5.0 out of 5 stars
Has everything a beginner needs to know and master AI
5.0 out of 5 stars
Find the problem you can solve
4.0 out of 5 stars
Great, but missing details that make it challenging for beginners
3.0 out of 5 stars
Minority Report
Top reviews from other countries
1.0 out of 5 stars
Kindle version is unreadable
5.0 out of 5 stars
Vermittelt interessante Erkenntnisse
Sie kann es sehr empfehlen,
5.0 out of 5 stars
El mejor libro de texto sobre inteligencia artificial
Los capítulos de Machine Learning y Deep Learning han sido actualizados de forma que incluyen los últimos avances, y al verificar los autores de cada capítulo se observa que los principales investigadores de cada área han colaborado.
Es un libro de texto que da para varias asignaturas trimestrales, y como consulta tiene un valor incalculable.
Lo recomiendo a todo estudiante avanzado de informática o matemáticas que quiera entrar en el mundo fascinante de la IA.
Ojo. Es un libro de texto, pesa un par de kilos (revisa las medidas porque es un libro grande)
El mejor libro de texto sobre inteligencia artificial
Los capítulos de Machine Learning y Deep Learning han sido actualizados de forma que incluyen los últimos avances, y al verificar los autores de cada capítulo se observa que los principales investigadores de cada área han colaborado.
Es un libro de texto que da para varias asignaturas trimestrales, y como consulta tiene un valor incalculable.
Lo recomiendo a todo estudiante avanzado de informática o matemáticas que quiera entrar en el mundo fascinante de la IA.
Ojo. Es un libro de texto, pesa un par de kilos (revisa las medidas porque es un libro grande)
Images in this review
5.0 out of 5 stars
The definitive and most comprehensive book on Artificial Intelligence.
- It is huge (1100+ pages).
- It is expensive, but nevertheless a very good investment. It is unlikely another edition of this book will be released for a long time (last version was published in 2010), and so this will stand the test of time for now.
- High quality hardback, with good binding (contrary to other reviews I have read).
- The pages are thin, but nevertheless high quality, with great use of colour on all pages, illustrations and diagrams.
- The book shipped from the US to the UK for delivery, and despite this still arrived pristine without any damage at all. I ordered through Amazon with Book Depository, who packaged it very carefully in a decent sized box with bubble wrap. I know some other sellers might not be so diligent, which isn't worth taking the chance with such an expensive book.
The definitive and most comprehensive book on Artificial Intelligence.
- It is huge (1100+ pages).
- It is expensive, but nevertheless a very good investment. It is unlikely another edition of this book will be released for a long time (last version was published in 2010), and so this will stand the test of time for now.
- High quality hardback, with good binding (contrary to other reviews I have read).
- The pages are thin, but nevertheless high quality, with great use of colour on all pages, illustrations and diagrams.
- The book shipped from the US to the UK for delivery, and despite this still arrived pristine without any damage at all. I ordered through Amazon with Book Depository, who packaged it very carefully in a decent sized box with bubble wrap. I know some other sellers might not be so diligent, which isn't worth taking the chance with such an expensive book.
Images in this review
1.0 out of 5 stars
التزمو