Artificial Intelligence Books

This book is an excellent reference book on current approaches to some of the foundational questions in artificial intelligence, philosophy of mind, and engineering of intelligent systems. The editors as amply demonstrated by their other publications, have the rare ability to interrelate an amazing diversity of perspectives on Artificial Intelligence - from symbolic methods to neural networks and evolutionary apporaches, to uncover the shared principles and common foundations, and show how different paradigms can be brought together in synergistic ways to advance our ability to design and analyze intelligent systems. This book includes chapters written by some of the leading experts in the field. The book should be especially useful as a reference to graduate students and researchers interested in artificial intelligence, machine learning, cognitive science, software agents, and philosophy of mind. It should also be useful as supplementary reading for a graduate or upper level undergraduate course in artificial intelligence or cognitive science, or as a primary text for a seminar course. All in all, a great book!

Symbolic (e.g., logic based) and connectionist (e.g., neural network) models are often viewed as separate, and perhaps incompatible approaches to artificial intelligence and cognitive modelling. Far too many young researchers (including myself) have been swept away by the exaggerated claims of each camp. This collection of chapters is an excellent medicine for such a malaise. The chapters in this collection demonstrate, with pursuasive theoretical and philosophical arguments as well as empirical data that symbolism and connectionism can, and perhaps should, be reconciled. At the time this book came out, this was not a popular position. However, recent developments in AI and cognitive science have more than vindicated the views expressed in this book. Now (6 years later), some of the chapters are a bit dated. However, I recommend the book strongly to those who are interested in foundational issues in artificial intelligence and cognitive science.

This book presents an exciting and challenging introduction to artificial intelligence and robotics for young people (probably ages 8-12). Jefferis describes many aspects of artificial intelligence, from the history of the field, to how computers work, robot sensors, gaming, neural networks, robot companions, and robot ethics. Each topic gets a 2-page spread, with at least as much space devoted to pictures as text. The text itself is clear and understandable, yet it doesn't come across as being dumbed down for kids. At the end of the book is an AI historical timeline, a glossary, and an index.

My 9 year old could not put this book down! Full of interesting photos and drawings, it teaches kids about current research in artificial intelligence (e.g., neural nets) and robotics. Amazing!

Penny Baillie-De Byl's ARTIFICIAL LIFE POSSIBILITIES: A STAR TREK PERSPECTIVE examines what the future might bring in creating exceptional artificial environments such as Star Trek has featured. Using Star Trek's innovative concepts as a foundation, artificial intelligence researcher Dr. Penmny Baillie-De Byl considers artificial life forms depicted in the TV series and the state of current technology to consider the potentials behind creating advanced AI systems.

The author has picked an interesting way to examine the state of the art in artificial intelligence. It is an interesting turn around. The originators of the TV series Star Trek looked at the research that was being done and extrapolated this research into their characters. In turn, these characters were looked at by researchers to see how their research might really be carried forward.

In this book, Dr. Penny Baillie-de Byl, an Australian university lecturer has in turned looked at the research being conducted and tied it back to the TV show characters. She looks at androids and at purely projection characters such as those that are generated on the holodeck.

The chapter I liked best was her discussion of the Turing Test, a test Alan Turing devised to determine if a machine could think - (what's think, what's a machine). Have we passed the test yet? Then again, I see some humans once in a while that I don't think could pass the test.

In an age when computers make ever greater inroads into our everyday lives, well may we ask: Do computers have intelligence? Are they living? Have free will? Exercise moral judgments? Stanley L. Jaki, historian and philosopher of science, deals with these and related questions in Brain, Mind and Computers, a thoroughly documented rebuttal of contemporary claims about the existence of, or possibility for, man-made minds. His method includes a meticulously documented survey of computer development, a review of the relevant result of brain reseach, and an evaluation of accomplishment of physicalist schools in psychology, symbolic logic, and linguistics, and a thorough critique of claims about artificial intelligence.

Comments on the first edition:

"Dr. Jaki's book is the most informed, pentetrating and lucidly written treatment of the subject that I have read anywhere." Robert A. Nisbet, Professor Emeritus, Columbia University.

"Certainly it is rewarding and refreshing to read such penetrating criticism of a field in which gratuitous theorizing and dogmatism are able to flourish bcause our scientific understanding is so small." Sir John C. Eccles, Nobel Laureate, 1963.

"This is a book fascinating in style as well as content...which every scientist should read." Eugene P. Wigner, Nobel Laureate, 1963.

"Dr. Jaki presents a sustained, well-informed, and persuasive argument for mind-body dualism...my own predilections are exactly opposite to Dr. Jaki's conclusions, but I welcome his challenge..." Herbert Feigl, University of Minnesota

[I meant to rate this with 4 stars.]

If you read enough of Fr. Jaki's works, or at least enough of the right ones, you see certain themes emerge time and again. One of the most important of those "Jakian" themes is the irreducible ontological gap between "the quantitative and the qualitative." Fr. Jaki explicitly cites an early source for this distinction as Aristotle (cf. Categories 16a). What makes physics the chief of natural sciences is its ability (sometimes envy-producing for other sciences) to isolate minute areas of material reality and explain them to an exhaustive quantitative degree. However, given the gap between quantities and qualities, this limits physics to quantitative concerns (when physics brings in literally meta-physical perspectives and assumptions, it makes proper use of the realm of qualitative reality). Given the nature of reality, you could call physics the supreme, because supremely limited, science. The disparity between quantities and qualities is the thesis of Fr. Jaki's first book on the history of science, **The Relevance of Physics** (TRP, 1966), its relevance being but the narrowly defined flip-side of its Irrelevance in many areas of life, an irrelevance acknowledged by many of physics' brightest lights.

The quantity-quality theme is also the driving force behind **Brain, Mind and Computers** (BMC). Indeed, Jaki mentions he originally intended to make BMC a closing chapter of TRP, but, upon reading M. Taube's **Computers and Common Sense**, he decided the cognitive/AI issue needed a lengthier, manifold treatment on its own. Ideally, then, BMC should be read in conjunction with, and perhaps only shortly after, TRP. BMC originally (ca. 1969) consisted of four chapters (each averaging 160 footnotes) and an epilogue, but in 1989 Fr. Jaki reissued BMC with a new fifth chapter (sort of like H. Dreyfus did with his **What Computers STILL Can't Do**, though Fr. Jaki thinks not very highly of Dreyfus's phenomenological arguments against strong AI), so be sure you get the newer paperback edition from Regnery.

Not only as a "Jakian" Catholic myself, but also as a believer in academic rigor -- one of Jaki's great strengths -- I am constantly miffed and surprised not to see this book cited in the indices or bibliographies of books dealing with the philosophy of mind and cognitive sciences. (An exception is D. Hofstadter's annotated bibliography in **Gödel, Escher, Bach**, but even then he brushes BMC aside as mere polemics, albeit with some "interesting" points ... yet he never engages those interesting points.) Certainly BMC is dated in terms of its contemporary analysis of AI. Even so, the gaps it fills in the historical record and the emphasis it lays on key issues -- such as 1) the futility of a physicalist reduction of human consciousness, 2) the important (rather Gödelian) discrepancies between human cognition and computerization (i.e., between language-as-understood and terms as formally describable), and 3) the crucial difference between computational results and intellection per se (i.e., the immateriality of thought per se). This last point deserves some elaboration. To borrow one of Fr. Jaki's own metaphors, just as two rivers may combine molecules when they converge but do not thereby perform addition, as a formal mental operation, so a computer may produce an algorithmic solution without thereby grasping the problem. The immateriality of intellection is understood by Fr. Jaki in terms of all words being universals and all meaningful discourse being predicated on methodical realism.

For these reasons alone, BMC should not be so consistently ignored by supposedly well read scholars in the field. The praise the book earned when it first appeared, coupled with the status of its author, should make BMC more prevalent in the discussion, even if only as a matter of academic thoroughness. BMC should remain especially significant in the AI/cog-sci debates since it is argued in tandem with TRP, a book no scholar of science can do without reading.

Of course, I am inclined to believe that, despite his accolades on a formally academic level, the priestly collar so proudly worn around Fr. Jaki's neck has led, even if unconsciously, to chronic disparagement of him on a personal level, moreso than some academics might care to admit.

Works that could profitably be read with BMC include:

M. Adler's **The Difference of Man and the Difference It Makes**
M. Adler's **Intellect: Mind over Matter**
J. Maritain's **The Degrees of Knowledge**
E. Gilson's **Linguistics and Philosophy**
E. Gilson's **Methodical Realism**
M. Taube's **Computers and Common Sense**
S. Jaki's "The Brain-Mind Unity" (Real View Books pamphlet)
S. Jaki's **The Relevance of Physics**
J. Ross's "Immaterial Aspects of Thought" (available via JSTOR)

Building Problem Solvers is a very hands-on introduction to AI systems in general, and truth-maintenance systems in particular. All the systems discussed come fully implemented. In this regard, the book can be viewed as an accessible and detailed discussion of the code. For this reason, the best way to read any chapter is to scan through it first, then, carefully go through it again with the source code in sight.

The authors were PhD students of Gerry Sussman at MIT. Thus, this book is a great way to learn about the classic AI systems and techniques devised and refined at the MIT AI lab.

If you get serious about the book, you'll want to try out a few of the exercises. I found that the exercises are invariably insightful, though I wish some included implementation hints, because elegant solutions are often far from obvious.

In short, I highly recommend this book if you're looking to build some problem solver using proven AI techniques.

This book is very rewarding to study and put into practice. I had a lot of fun immersing myself in the concepts and systems developed in this book. Using the techniques of this book as a base, we've implemented BioHacker, a debugger for metabolic networks.

This book takes you on a journey through the techniques that have been used to build problem solvers, ranging from classical search techniques to Assumptive Truth Maintenance. It is very easy to read, despite the high level of technical detail. The LISP code that accompanies the book is well documented, easy to understand, and it works. It is a must have for anyone who truly wants to understand problem solving techniques.

Don't you know Cellular Neural Networks? Read this book, even if you are not a specialist. I like it a lot, it's very interesting. I had a course project on CNN for my master in electronic engineering but I never found a so clear book in describing principles, applications and circuit realizations about CNN. Don't be scared if you can find some high level theoretical contents, they are for specialist in that field only, but they are not necessary for novices to know of CNN. Excellent!

I definitely enjoyed reading this book. I found it extremely clear and interesting. I am an electrical engineer but I am certainly not a specialist of this field. However, I catched up with the subject very quickly. And it is amazing to see how many applications these CNN have. As a matter of fact some of the discrete component circuits here described can even be realized by an electronics hobbyst like me. And I plan to do my own experiments. I admit I wasn't able to follow some parts with high level theoretical contents (some math definitely goes beyond my understanding). Nevertheless I found it very appealing. In summary I would encourage even the novices to check it out. I like it a lot.

This book is one of the pioneer works on the subject, as its publication date shows clearly. Despite its age, it remains as one of the most fascinating introductions to computer chess, and most of the ideas it presents are still valid. Its multiple authors cover all aspects of chess playing, from technical expositions of some of the best programs of that time, to physiological and psychological considerations.

In "A brief history of computer chess tournaments: 1970-1975", we are introduced to the atmosphere of the early tournaments, the diverse friendly matches between US and USSR chess computers, and several US and international championships, with many of the most interesting games fully commented and analyzed.

The next chapter, "Human chess skill" focuses in how does a human player select a move in the game of chess, the role of perception, the search mechanism, visualization, as well as other tipically human aspects such as motivation. Several tests applied to human players ranging from novices to grandmasters are presented and discussed.

After that introspective look at we humans, and our not-so-well understood thought processes, "An introduction to computer chess" begin to shift the focus to the computer, including such basic topics as how to represent the chess board, the moves, the status, how to generate the legal moves, search strategies, position evaluation, so that by the end of the chapter, all necessary foundations are well stablished for the rest of the book.

With Chapter 4, "Chess 4.5 - The Northwestern University chess progam" we begin the most technical part of the book. Here, authors David J. Slate and Lawrence R. Atkin show us with great style the internal workings of their famous chess program, many times world champion, and the one mostly used against IM David Levy for the famous Levy's bet. The details are sufficient to help a lot anyone contemplating the possibility of writing his/her own chess program. Modestly, the authors assume the limitations of their creature, and offer good advice on how it can be incrementally improved.

Chapter 5, "PEASANT: An endgame program for kings and pawns" provides yet another close scrutiny of a chess program, though this time with the important novelty that it is an specialized chess program, one specifically designed for a certain class of very frequent endgames. Monroe Newborn, its author, fully describes the inner workings, and most importantly, produces a set of tests for his program, with commented results.

The next chapter, "Plans, goals, and search strategies for the selection of a move in chess" tries to center on how do human players select good chess moves when having just a few seconds to consider the position (i.e: blitz chess), and then introduces a chess program specifically designed to play speed chess, without recourse to tree searching. This quite intriguing approach more closely mimics the human behaviour, to the point of even producing the same kind of erroneous moves a human player would play at blitz speeds.

As an alternative to the standard alpha-beta search techniques, Larry R. Harris introduces us to "The heuristic search: An alternative to the alpha-beta minimax procedure", where it presents what it considers important pitfalls of that search strategy, fully commented with specific examples, and proposes a new paradigm that addresses each and everyone of them from the start, thus truly directing the search in an intelligent way, as opposed to brute force, so that each aspect of the position can be ascertained as soon as possible, before going to other places in the search tree.

After these mostly technical chapters, in "Man and machine: Chess achievements and chess thinking", professor Eliot Hearst, a member of the Psychology Department at Indiana University, evaluates the present status of computer chess from the perspective of someone very knowledgeable with the game, as he is a rather skilled chess player and columnist. He includes many good practical examples, to make his points even clearer.

The book closes with a number of games played by Chess 4.5 and 4.6 in competitions during 1976, 1977, and 1978, that show a remarkable improvement on the rather pessimistic forecastings most experts agreed upon at that time.

This book shows the state of the art at the end of the 1970's. Though there have been huge changes since then, and tremendous gaps in our understanding have been filled, I still can't recommend this book highly enough. This book was published at the time when the best programs changed over from selective search to brute force. Nowadays we know that brute force is the way to go, but at that time even programmers who were winning tournaments using brute force techniques had little faith in their ultimate viability. The authors's speculation about the roles of search and evaluation is very interesting from the historical perspective.

One chapter of this book is worth the entire price. Slate and Atkins describe Chess 4.5 in one chapter. That chapter remains to this day the best description of an "attack-table" chess engine ever written, though you will need some additional reading to create a modern program on that basis.

This book is to the field of Artificial Intelligence what Darwin's Origin of the Species was to the idea of evolution. A historical document whose ideas are still being explored, it showcases all of the seminal papers and thinkers that began research in the area. These primary sources are invaluable for understanding the history of the AI, which is celebrating its 50th anniversary at the Boston AAAI conference as I type. Many of the very same authors and their students are presenting, still leaders in the field.

It looks like Alan Mathison Turing's suggestion stands the test of time: come up with a program that apes a child's mind and then put the computer through school. That may be what it takes to bring COMPUTERS AND THOUGHT together in the first real thinking machine. Proving theorems is good training in problem solving, which we learn along with verbal training, concept formulation, decision making under uncertainty, and social behavior. The first high order artificial intelligence, the logic theorist, finds proofs for theorems but gets into trouble when answers are beyond rote recalling of the table of integrals. We pull in intuition, transformations and tricks beyond elementary calculus, but the computer problem solves according to the way it was designed. So if computers could be designed to learn from routines, they would recognize patterns in proofsearch procedure and then strategize to repeat the patterns in later proofs. At the last page I felt energized from having gotten through such a challenging, excellently organized, and fascinating book. Editors Edward A Feigenbaum and Julian Feldman make a major contribution to information on artificial intelligence. So their book fits in with George Dyson's DARWIN AMONG THE MACHINES, Stan Franklin's ARTIFICIAL MINDS, David Freedman's BRAINMAKERS, John Haugeland's ARTIFICIAL INTELLIGENCE, or Clifford A Pickover's COMPUTERS AND THE IMAGINATION.

I foudn this book inspiring. It develops on the concept of "predispositioning", taking it from the initial development in chess, to all aspects of life. Predisposition is a concept that can be used as a guide to life, as the author presents cases in which he used it to improve his position (an analogy to chess).

The book brilliantly extrapolates from the premise of chaos theory and indeterminism. A great work!!

Especially given that this book is in fact a doctoral dissertation, it's extremely easy to read. This is not to say that it is written for children, but rather, the author has used language well to convey concepts rather than to confuse and sound stuffy.

The book states the limitations of the traditional Von Neumann computer architecture (which by and large we are still stuck with today) and then goes on to explain how an entirely different approach with many processors could work.

This reference describes a computer architecture containing thousands of processor/memory cells that can be connected together by software, and the rational behind this architecture. It is easy to read, and is useful in providing the general reader with a feel for large multiple processor computation, in particular an architecture well suited for semantic network marker propagation.