Confabulation Theory: The Mechanism of Thought (Hardcover)

~ (Author)

Editorial Reviews

Product Description

Confabulation theory offers the first complete detailed explanation of the mechanism of cognition, i.e., thinking, an essential information processing capability of all enbrained Earth animals (bees, octopi, trout, ravens, humans, et al.). Concentrating on the human case, this book offers an hypothesis for the neuronal implementation of cognition, and explores the mathematics and methods of application of its mechanism. Thinking turns out to be starkly alien in comparison with all known technological approaches to information processing. While probably not yet scientifically testable, confabulation theory seems consistent with the facts of neuroscience. Beyond science, any complete detailed explanation of cognition can be investigated by applying it technologically. Multiple experiments of this nature are described in this book in complete detail. The results suggest that confabulation theory can provide the universal platform for building intelligent machines. In short, this book explains how thinking works and establishes the foundation for building machines that think.

With two DVDs of courseware, this book is suitable both as a text for professional self-study and for graduate and advanced undergraduate courses in neuroscience, computational intelligence, cognitive science, linguistics and psychology, the prerequisites being elementary mathematics and neuroscience. Because of the theory’s implications for philosophy, education, medicine, anthropology and social science, this book will also be of interest to scientists in those domains.

About the Author

Robert Hecht-Nielsen was made a Fellow of the IEEE in 1997 for leadership in practical applications of neural network technology. He was a pioneer in the development of neural networks and authored the first textbook on the subject, Neurocomputing (1989). He has been a member of the UCSD faculty since 1986.

Product Details

• Hardcover: 245 pages
• Publisher: Springer; 1 edition (August 24, 2007)
• Language: English
• ISBN-10: 3540496033
• ISBN-13: 978-3540496038
• Product Dimensions: 9.4 x 6.3 x 0.7 inches
• Shipping Weight: 1.3 pounds (View shipping rates and policies)
• Average Customer Review: 4.0 out of 5 stars  See all reviews (3 customer reviews)
• Amazon.com Sales Rank: #1,322,410 in Books (See Bestsellers in Books)

Customer Reviews

Most Helpful Customer Reviews

6 of 6 people found the following review helpful:

4.0 out of 5 stars Outside the box, November 18, 2007

By	Jim Lewis (NJ USA) - See all my reviews (REAL NAME)

Amazon Verified Purchase

I heard about confabulation theory from an AGI video and looked forward to this book. The author presents a unique and relatively simple theory about the basic algorithm of the brain. Although the theory is incomplete, the big benefit of the book is presenting a thought provoking new view - and one based (I am not in a position to assess how accurately) on a substantial body of neuroscience. There is substantial redundancy resulting from use of text from the author's scientific papers - I kept asking myself, "didn't I already read this"? And although I'm in the KISS camp I am doubtful that the core algorithm can be quite as simple as confabulation theory supposes. Yet at the core are some fascinating ideas that present fuel for a much needed new direction in thinking about how to approach AGI, and sufficient concreteness from which to develop code experiments for AGI implementations.

3 of 3 people found the following review helpful:

3.0 out of 5 stars Well-written, but repetitive and overstated, April 26, 2008

By	G. Snider (Palo Alto, CA USA) - See all my reviews (REAL NAME)

"Confabulation Theory" is not a textbook but a collection of papers by the author plus two talks on DVDs. The writing is very clear and the talks are excellent, taking you through the theory one step at a time. Hecht-Nielsen is a great teacher. But the papers are repetitive, covering the same basics over and over again, while skimming over the key topics of self-organization within modules and implementation of the winner-take-all networks. The book spends too much time on sentence confabulation experiments (which appear to be similar in approach to statistical language translation) and not enough on the applications sketched out in the last two chapters (vision and continuous speech segmentation).

Hecht-Nielsen has an unfortunate habit of overstatement which weakens his presentation. In the text and the videos he repeatedly points out how "starkly alien" confabulation theory is, and how it might revolutionize "psychology, education, philosophy, psychiatry, medicine (both human and veterinary), law, and theology." Really? If you're unfamiliar with neuroscience, the massive parallelism and nonlinear dynamics might seem alien, but this is old hat in the field. His "Fundamental Theorem of Cognition" goes over the edge, though, seemingly suggesting that he has "solved" the problem of cognition once and for all; the "fundamental theorem of confabulation" would have been a more appropriate and less annoying name.

There are a lot of interesting ideas here, and I finished the book wanting to know more. I hope that someday Hecht-Nielsen writes a book with greater depth in implementation issues and more breadth on applications. The papers making up the core of the theory are available on the Web, so the book is worth buying only for the convenience of having everything bound together.

0 of 2 people found the following review helpful:

5.0 out of 5 stars Enthusiastic presentation, strong experimental results, October 3, 2008

By	Donald C. Wunsch II (Rolla, MO USA) - See all my reviews

Full disclosure -- I've known Robert Hecht-Nielsen since 1986, when I took his short course on neural networks at UCSD. He's a respected friend and colleague.

Robert's passion for the field and his research has resulted in an enthusiastic presentation style that apparently struck one of the other reviewers as "over the top". But the plus side is that this is a contagious passion. This is particularly true given his outstanding results, particularly in sentence completion, next sentence generation, and object segmentation. He is a talented researcher with substantial neuroengineering and design experience, plus a broad and deep knowledge of the field, as well as a strong research team, so these advantages have to be taken into consideration when evaluating his results. That said, the results speak for themselves and make this book a worthwhile read.