Artificial Intelligence Books

You simply cannot choose a better overview of intelligent agent modeling. Wooldridge brings the reader from an extremely non-threatening, non-technical treatment of the basics to formal modeling of individual agent properties and multi-agent interaction using (essentially) constructions of formal logic.
This book is clearly intended to communicate a particular theory in a very effective manner. However, it does not (nor does it pretend to) provide a path to implementing any of the theory, so dilettantes beware.

An excellent treatment of the formal modelling of what Mike calls Rational Agents and others have (more provocatively) usually termed Intelligent Agents. What sets this apart from other books on the subject are Mike's refreshingly direct, no-nonsense writing, and his recognition that his audience may not have all of the formal tools at their fingertips. The Appendix "Formal Foundations" would be a good choice for anyone who wanted to understand modal logics, especially as they apply to time and intention. Even software engineers who would rather hack code than grapple with formalism will find this accesible, stimulating and useful.

I give this book 5 stars. It is a must have book, very well written. It has good balance between rigorous theory and authors reasonong regarding the subject.
I'm not a beginner in this field, and still I found a lot of interesting ideas, that can help not only to improve quality of the net, but also make you see "bigger picture".

"Recurrent Neural Networks for Prediction: Learning Algorithms,
Architectures and Stability," approaches the field of recurrent neural networks from both a practical and a theoretical perspective. Starting from the fundamentals, where unexpected insights are offered even at the level of the dynamical richness of simple neurons, the authors describe many existing algorithms and gradually introduce novel ones. The latter are convicingly shown to yield better prediction performances than traditional approaches, when applied to real-world data. They also dedicate a considerable amount of time on the (practical) issue of nonlinearity analysis of time series, which is or should be, indeed, the cradle of all proper modelling and/or filtering solutions: nonlinearity should be assessed prior to choosing the appropriate model and/or filters, since linear ones are to be preferred if sufficient for the problem. I would recommend this book to any researcher who is active in the field of recurrent neural networks and time series analysis, but also to researchers who are new in the field, since the book offers an extensive overview of the current state-of-the-art approaches.

We used this book for a month for a course in distributed AI. This books provides a comprehensible structure to multiagent negotiations by identifying three different kinds of domains (task oriented, state oriented, and worth oriented), and describe negotiation protocols and give their important properties. Wherever needed, the authors make the connection to game theory. This is a solid base to recent papers and books on multiagent negotiations, and also to build your own MAS using negotiation protocols of varying levels of complexities.

We used this book for a month for a course in distributed AI. This books provide a comprehensible structure to multiagent negotiations by identifying three different kinds of domains (task oriented, state oriented, and worth oriented), and describe negotiation protocols and give their important properties. Wherever needed, the authors make the connection to game theory. This is a solid base to recent papers and books on multiagent negotiations, and also to build your own MAS using negotiation protocols of varying levels of complexities.

I spent some time looking for a simple bedside aggregation of the various topics associated with the psychology of decision making and the various perceptual biases, without finding much. Most of the books are excellent; but, aside from this one (and Jon Baron's) they are usually compilation of original research. I like to have a readable consolidation of the material not far from my figertips. I was lucky to have found this book, which provides a wonderful and comprehensive coverage of the topics.
It is limpid, precise, illustrative, showing a wonderful clarity of mind.

Now the bad news. The author passed away recently at the age of 48.

The book--although poorly bound--is written very nicely. Kunda has a wonderful way of presenting the material objectively and respectfully while still engaging the reader. There are numerous, minor typos but this does not distract the reader terribly. A natural progression is found in Kunda's style, where she presents a topic, draws on some familiar day-to-day experiences, and presents the experiments and research that support it. She explains contradictions and allows the reader to make his/her own conclusions and inferences concerning topics like hot cognition, influences of mood on memory, subliminal processing, false-memories, and influences of culture on cognition. She, herself, did extensive work in the area of cognitive science and since her passing, will be missed by many as her work was incredibly innovative and interesting.

The book reads quickly and will be enjoyed by most. The only thing I wish was different about this book was its size (there is so much left over white space on each page--it's unnecessary). Most people with some experience or interest in cognitive science should find this book fascinating. Do not expect an in-depth or even moderate assessment of the biological or neurological underpinnings of cognition. This book explains cognition by reviewing social and cognitive experiments.

Social Fuzziology explores the fuzziness inherent in what we know about ourselves and society, about the sources and nature of our experience, of our thoughts and feelings, drives for understanding and urges to create and realise our potential. This kind of fuzziness is at the very essence of human existence, therefore it affects any field of people's activity, be it engineering design of intelligent systems or social inquiry into enigmas of life. The book draws on the full range of the social sciences, including some of the most up-to-date work on escalating complexities of post modern life, e.g. new information technologies in the global world, new forms of consciousness and identity, new relationships between humans, machines, nature and environment. The volume is interdisciplinary, combining expertise from fuzzy logic and the postmodern social sciences to outline, synthesise and implement the challenging concepts and methods of Social Fuzziology. It gives readers from the social sciences a comprehensive, exciting and practical guide for applying the insights and approaches of fuzziology, while bringing a rich body of social problems and concepts to the agenda for researchers in fuzzy logic and related fields.

Social Fuzziology explores the fuzziness inherent in what we know about ourselves and society, about the sources and nature of our experience, of our thoughts and feelings, drives for understanding and urges to create and realise our potential. This kind of fuzziness is at the very essence of human existence, therefore it affects any field of people's activity, be it engineering design of intelligent systems or social inquiry into enigmas of life. The book draws on the full range of the social sciences, including some of the most up-to-date work on escalating complexities of post modern life, e.g. new information technologies in the global world, new forms of consciousness and identity, new relationships between humans, machines, nature and environment. The volume is interdisciplinary, combining expertise from fuzzy logic and the postmodern social sciences to outline, synthesise and implement the challenging concepts and methods of Social Fuzziology. It gives readers from the social sciences a comprehensive, exciting and practical guide for applying the insights and approaches of fuzziology, while bringing a rich body of social problems and concepts to the agenda for researchers in fuzzy logic and related fields.

I was a student of Murray when I was doing my master's in AI. He taught us a course titled 'Learning and Cognitive Robotics', which covered a classical 'logic-based' approach to robotics, starting from the 'Shakey' days to Event Calculus.

Much of the lecture notes we covered in the said course went into the book, which to my mind provides a comprehensive coverage to logic-based AI, its challenges and possible directions.

In classical AI, we assume that 'correct reasoning on correct representation' preceedes intelligence. The book starts from this 'first principle' and takes one through a wonderful journey, leading to an almost total solution to the Frame Problem.

In my mind, the book is not for naive readers. One has to have a fairy good understanding of relevant Computer Science/AI concepts to comprehend the material.

I was a student of Murray when I was doing my master's in AI. He taught us a course titled 'Learning and Cognitive Robotics', which covered a classical 'logic-based' approach to robotics, starting from the 'Shakey' days to Event Calculus.

Much of the lecture notes from the said course went into the book, which to my mind provides a comprehensive coverage to logic-based AI, its challenges and possible directions.

In classical AI, we assume that 'correct reasoning on correct representation' preceedes intelligence. The book starts from this 'first principle' and takes one through a wonderful journey, leading to an almost total solution to the Frame Problem, one of the biggest stumbling blocks for logic-based AI.

In my mind, the book is not for naive readers. One has to have a fairly good understanding of relevant Computer Science/AI concepts to comprehend the material.

For serious readers of AI, I would like to recommend another fantastic piece of work: Agent_Oriented Programming: From Prolog to Guarded Definite Clauses by Ringwood and Huntbach.

Enjoy!

Overall - Excellent Read!

For a little background, I used this book as an introduction to cognitive science as part of an undergraduate, self-study, artificial intelligence course. As a result of this book, I now have a greater appreciation for what AI is, I know why it is generally looked down upon by those outside of the field, and I still got a good overview of cognitive science.

While this book is a little dated (the interviewees talk a lot about connectionism), it is an excellent way to understand the history of AI and cognitive science. It is also quite nice to see some of the participants criticize each other presenting two side to an argument (particularly Dreyfus and Newell on the progress & future of AI).

As a result of this book, I have a much better appreciation for cognitive scientists in general and AI researchers specifically. I think this should be on the short list for any aspiring AI researcher.

It was an thoroughly interesting read. The book touches on everything cog sci.

I am new to speech processing and this book was my starter for this subject. It is a very good beginning, but still very indepth leaving information for later studies. It is structured very logical and the extensive literature and web references are very helpful

The book provides a solid introduction to all aspects of speech communication, including human-human communication (production, hearing and perception), physical aspects of spoken language (speech analysis), speech coding and enhancement for efficient storage and transmission, as well as speech recognition and synthesis, and speaker recognition.
Each chapter explains the basic principles and is suitable for readers with some background in signal processing, computer science or phonetics. However, beyond the basic principles, the book contains an extensive survey of the state of the art in the different areas and can serve as an up-to-date reference book for many areas of speech communication, and often also provides a brief discussion of new and controversial research areas. Each chapter contains a comprehensive lists of references, and the book provides pointers to web resources in all areas of speech processing.

Diane Forsythe was trained in both anthropology and artificial intelligence, and her work offers insights to both of these realms. It's a pleasure to have her work collected and published (posthumously) in this volume. Many of the issues which Diana Forsythe confronted implicitly and explicitly in her work are issues which are of ongoing concern for all anthropologists. Her concerns included the responsibilities, opportunities, and limitations that come with funding sources, and, not unrelated, the numerous challenges associated with `studying up.' Finally, Forsythe struggled with the difficulty of balancing her participant observation with social action.

Forsythe argued that `studying up' disrupts the `traditional fieldwork narrative,' and at the same time calls into question all that is familiar ground for anthropologists from an ethical and methodological perspective. Often, said Forsythe, this type of participation requires anthropologists to be experts not only in anthropology, but also to some extent in their area of inquiry. As a result, we may have similar skill sets and jobs to those we study, but we may also be competing against them for funding and/or workplace resources. This twist poses a complex ethical dilemma for which anthropologists do not readily have the answers. In addition, informants can read what is being written about them, which results in a new accountability and criticism as part of the research process.

The Medical Anthropology Quarterly article provides one example of the many challenges she encountered in her work. In it, Forsythe argues that computer programs are a product of the culture from which they emerge, and that her intent is to describe how this occurs. She says that the program which was designed to aid migraine sufferers espouses the physicians point of view, and it may in fact reinforce the divide in the doctor-patient relationship rather than improve it. Rather than developing the questionnaire from a social sciences perspective, she describes a team dynamic where a programmer understands this task to be a logical one rather than a social one, and so takes this on as his primary task. As a result, the questions in the tool are structured in such a way that they do not make room for issues like violence against women, which is potentially a major factor in the case of female migraine sufferers. She also says that the composition of the team included no nurses, and, although members of the team were migraine sufferers, they nonetheless spoke with their professional perspective throughout the process, and never with the perspective of a patient. She goes on to describe the nature of the program that was created, saying that it really only answers questions that patients would normally ask in the office, and that, as a result, it did not address any of the needs of the patient that were not addressed in clinic visits. Ultimately, then, the system reinforces the doctor's diagnosis, and provides little to no new information or insights to the patient. Through her description, she offers a set of issues and concerns which change the perception of the project and provide insight into the tool and its limitations.

Many anthropologists have attempted to answer the question of how to be sufficiently engaged and disengaged simultaneously. How can we - or can we - be agents of change in the world we study? Forsythe illustrates the challenges of finding paid work in non-academic anthropology, the complexities of studying up which may accompany that work, and the difficulties with being engaged in critical practice at the same time.

I first met Diana Forsythe at the 1992 annual meeting of the American Medical Informatics Association. She was one of the few people I'd met who shared my interests in Information Technology's potential to empower patients.

Diana told me about a recent study she'd done in which the physicians caring for migraine patients had little idea of their patients' chief concerns and thus did a woefully inadequate job of dealing with their information needs. Most of the patients had never even asked their doctors the question that concerned them most during the long painful attacks they sometimes experienced-"could I die of this?" Since the doctors all knew that no one dies of migraines, they considered this a trivial and irrelevant concern and somehow discouraged questions about it.

Over the months that followed, Diana and I had several wonderful long conversations about the need to understand and rethink some of the tacit assumptions of current healthcare system as a necessary prelude to attempting to computerize it. And so, like many of her other friends and colleagues who had been impressed by her brilliance and touched by her intellectual generosity, I was heartsick, some years later, when I learned that she had drowned in a raging river while backpacking in Alaska.

A group of her concerned colleagues, worried that her remarkable articles and talks might be lost, banded together and convinced Stanford University Press to publish a volume containing her collected works. The book, long in preparation, has just been published. It is entitled: "Studying Those Who Study Us: An Anthropologist in the World of Artificial Intelligence, by Diana E. Forsythe (Stanford University Press, Stanford, CA, 2001 ISBN 0-8047-4203-0)

If you are a clinician or medical informatics professional interested in developing effective IT systems for patients, I strongly suggest that you order this book today. And when it comes, turn directly to Chapter Seven: "New Bottles, Old Wine: Hidden Cultural Assumptions in a Computerized Explanation System for Migraine Sufferers." This may be the most insightful paper ever written on the subtle difficulties that typically arise when doctors try to build IT systems designed to be used by patients.

To build effective online health systems for end-users one must combine the knowledge of a medical professional, the skills of a programmer/developer, the perspective of a medical anthropologist, and the wisdom of Solomon. And since Solomon is not currently available, an insightful social scientist like Diana-who can help us see our current healthcare practices from a 'man-from-mars' perspective-can offer invaluable insights.

Selected quotes & summaries:

"Both builders and users of [CHI] systems tend to think of them simply as technical tools or problem-solving aids, assuming them to be value-free. However, observation of the system-building process reveals that this is not the case: the reasoning embedded in such systems reflects cultural values and disciplinary assumptions, including assumptions about the everyday world of medicine."

"The system-builders systematically excluded the possibility that many of the migraines in female patients might be related to domestic violence, even though the ethnographers' interviews suggested that this was indeed the case."

I laughed, I cried, it changed my life!

Since AI researchers are generally engineers, they historically did what engineers do: they broke up the mind in very clear-cut divisions, one for the perception of the things out there in the world, and another, symbolically, to do "abstract cogitation".

For deep reasons, this was an invalid move, but only a few could see it. Robert French surely could, for his highly original tabletop project exhibits some of the best insights in Artificial Intelligence ever.

AI is still so much pervaded with the wrong ideas that this book will need to take some time to make its definitive mark on the history of the field.

If genuine understanding is ever to be built into a machine, understanding of the kind that Searle's gang will be forever denying, then it will come from an architecture similar to that proposed in this book.

Then again, I could turn out to be wrong. But let us let History decide on this issue.