New Book: What is Thought? Eric B. Baum
MIT Press 495p Best price right now is at Barnesandnoble.com $32, with free shipping. To buy this book: Barnes and Noble.com: http://search.barnesandnoble.com/booksearch/isbnInquiry.asp?userid=2WI405VPJU&isbn=0262025485&itm=17 Amazon: http://www.amazon.com/exec/obidos/tg/detail/-/0262025485/qid=1074532277/sr=1-3/ref=sr_1_3/002-6265544-0286451?v=glance&s=books MIT Press: http://mitpress.mit.edu/catalog/item/default.asp?sid=AF8A6531-E5E9-4710-A781-CA47C6B64621&ttype=2&tid=9978 Can the strong AI/ Turing picture be extended to a plausible model of all aspects of mind, such as understanding, creativity, language, reasoning, learning, and consciousness? I propose a candidate extension that is consistent with extensive data from a variety of fields and makes empirical predictions. In my view meaning is the computational exploitation of the compact underlying structure of the world, and mind is execution of an evolved program that is all about meaning. I extrapolate computer science research on Occam's Razor to argue that meaning results from finding a compact enough program behaving effectively in the world; such a program can only be compact by virtue of code reuse, factoring into interacting modules that capture real concepts and are reused metaphorically. For a variety of reasons, including arguments based on complexity theory, developmental biology, evolutionary programming, ethology, and simple inspection, this compact Occam program is most naturally seen to be in the DNA, rather than the brain. Learning and reasoning are then fast and almost automatic because they are constrained by the DNA programming to deal only with meaningful quantities. Evolution itself is argued to exploit meaning in related ways. Words are labels for meaningful computational modules. Using the abilility to pass along programs through speech, humans have made cumulative progress in constructing useful computational modules built on top of the ones supplied by evolution. The difference between human and chimp intelligence is largely in this additional programming, and thus can be regarded as due to better nurturing. The many aspects of consciousness are also very naturally and consistently understood in this context. For example, although the mind is a distributed program composed of many modules, the self emerges naturally as a reification of the interest of the genes. Qualia (the sense of experience of sensations such as pain or redness) have exactly the appropriate nature and meaning that evolution coded in the DNA so that the compact program behaves effectively. This book is highly relevant to the AGI agenda in many ways, surveying much of the progress of AI with an eye toward why it has fallen short of general intelligence, proposing a theory of how mind achieves general intelligence, and discussing what steps would be useful to achieve general intelligence computationally. Because I argue that most of the computational work in producing mind was done by evolution, rather than in our brains during our lifetimes, I view the computational complexity of building an AGI as significantly more challenging than other authors. However, I also present results of evolutionary programming experiments in the Hayek model, where we have succeeded in evolving, from random code, programs that solve classes of difficult planning problems in ways that seem to give intuition into how a program can achieve understanding. We were able to achieve these results because an analysis of evolutionary dynamics suggested mechanisms that greatly speed such evolution. --------------------------------------- >From the back cover: "This book is the deepest, and at the same time the most commonsensical, approach to the problem of mind and thought that I have read. The approach is from the point of view of computer science, yet Baum has no illusions about the progress which has been made within that field. He presents the many technical advances which have been made -- the book will be enormously useful for this aspect alone -- but refuses to play down their glaring inadequacies. He also presents a road map for getting further and makes the case that many of the apparently 'deep' philosophical problems such as free will may simply evaporate when one gets closer to real understanding." --Philip W. Anderson, Joseph Henry Professor of Physics, Princeton University, 1977 Nobel Laureate in Physics "Eric Baum's book is a remarkable achievement. He presents a novel thesis -- that the mind is a program whose components are semantically meaningful modules -- and explores it with a rich array of evidence drawn from a variety of fields. Baum's argument depends on much of the intellectual core of computer science, and as a result the book can also serve as a short course in computer science for non-specialists. To top it off, *What is Thought?* is beautifully written and will be at least as clear and accessible to the intelligent lay public as *Scientific American*." --David Waltz, Director, Center for Computational Learning Systems, Columbia University "What's great about this book is the detailed way in which Baum shows the explanatory power of a few ideas, such as compression of information, the mind and DNA as computer programs, and various concepts in computer science and learning theory such as simplicity, recursion, and position evaluation. *What is Thought?* is a terrific book, and I hope it gets the wide readership it deserves." --Gilbert Harman, Department of Philosophy, Princeton University "There is no problem more important, or more daunting, than discovering the structure and processes behind human thought. *What is Thought?* is an important step towards finding the answer. A concise summary of the progress and pitfalls to date gives the reader the context necessary to appreciate Baum's important insights into the nature of cognition." --Nathan Myhrvold, Managing Director, Intellectual Ventures, and former Chief Technology Officer, Microsoft Eric B. Baum has held positions at the University of California at Berkeley, Caltech, MIT, Princeton, and the NEC Research Institute. He holds a BA and MA from Harvard and a PhD in physics from Princeton. He is currently developing algorithms based on Machine Learning and Bayesian Reasoning to found a hedge fund. ------- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
