- Series: Oxford Series on Cognitive Models and Architectures
- Hardcover: 480 pages
- Publisher: Oxford University Press; 1 edition (June 13, 2013)
- Language: English
- ISBN-10: 0199794545
- ISBN-13: 978-0199794546
- Product Dimensions: 10.1 x 1.5 x 7.2 inches
- Shipping Weight: 2.6 pounds (View shipping rates and policies)
- Average Customer Review: 14 customer reviews
- Amazon Best Sellers Rank: #1,385,857 in Books (See Top 100 in Books)
Enter your mobile number or email address below and we'll send you a link to download the free Kindle App. Then you can start reading Kindle books on your smartphone, tablet, or computer - no Kindle device required.
To get the free app, enter your mobile phone number.
How to Build a Brain: A Neural Architecture for Biological Cognition (Oxford Series on Cognitive Models and Architectures) 1st Edition
Use the Amazon App to scan ISBNs and compare prices.
Frequently bought together
Customers who viewed this item also viewed
"[T]his book makes the complex nature of the brain clearer for laymen as well as neuroscientists. This is a definitive and pioneering work on the study of the human brain..." -- Biz India
About the Author
Chris Eliasmith is Canada Research Chair in Theoretical Neuroscience at the University of Waterloo
Top customer reviews
There was a problem filtering reviews right now. Please try again later.
There has been a long, long pause between the late-1980s work on neural networks and AI and some grand "unified theory" to connect the two into some reasonable explanation of how the human mind works. This book may not have *the* answer, but it has *an* answer, which is more than anyone else has proposed to date (apart from a slim line of research that is well and thoroughly cited throughout the book), and it certainly points the way forward. I picked up this book on a whim after seeing it mentioned in a quick citation in Dan Dennett's "From Bacteria to Bach and Back", and I must say that I have not enjoyed catching up on cognitive science research so much since reading Rumelhart and McClelland (apart from the book "Spikes", which is so breathtakingly expensive for such a slim volume that I have not obtained a copy for myself). "How to Build a Brain" is, like Parallel Distributed Processing, divided into two main parts, the first to describe the underlying mechanisms, equations, algorithms, and structures; and the second to describe a series of experiments working their way up from the simple to the surprisingly complex.
A shorter third part gives an overview of similar past efforts and compares them to the work of Eliasmith and his students/colleagues. The book contains pointers to the group's website where many (all?) of the software tools used in this research can be found, downloaded, and played with.
The book doubles as an "owners manual" for the author's Nengo neuro sim program (meaning Neural ENgineering Objects, not the Japanese era). Videos are available on how the book and system integrate-- check out section 1.5 in Amazon's generous preview peek for the website address. While you're at it, there is a LOT of detail given in this worthwhile preview that will immediately help you determine if this is the book for you.
I design domain specific languages for robotics and am always interested in new and different views of neural vs. semiconductor cognitive sims. Unfortunately to be honest if not crass, most of the books out there today are big on speculation and light on anything you can model. That's where this book shines from my narrow frame-- you can build your own versions of the authors premises and "try" them on what has become the biggest, most active and accessed "online brain sim" on the planet!
Do you have an interest in sims of semantics or perception? My slice is more on "translation" from the compiler side, but this book is wonderful in its coverage of practical aspects of semantics, perception, memory, what the authors call "semantic pointers" and of course the must mention topics today of plasticity and fluid intelligence.
I'll wait for a Neurological researcher to opine on this from the biological side, but for pure "translational" delight between brain and robotic mechanisms, this book can't be beat. It really is the only up to date text on the topic, as extensive brain sims are mostly covered in journals. You should know that the majority of the book IS meant to present a new and coherent theory of brain architecture, but the amount of detail, history, where we have been and where we are going the author covers is worth the price of the book itself, and of course is architecture independent.
As an EE I love math, but if you're math adverse, don't dismiss this, as the author has relegated the most complex math to the appendices. (You can't read neuro without running into differential equations due to the importance of dynamical systems in the models. But from an EE viewpoint, asp and dsp also involve time vs. frequency and we use similar Fourier etc. methods for different reasons on the robotics side).
Very readable, enlightening, and frankly fun, since you really can get hands on with the sims. Highly recommended, but suggest you take advantage of the Amazon preview before investing. Hats off to authors and publishers who have enough pride and confidence in the value of their work to let us have a significant preview like this.
Emailer answer: "What does Eliasmith think of Newell?" A. He mentions him throughout, with respect, but with the objective of bringing Newell's architectures closer to biological underpinnings. SOAR and other production systems are covered, but Chris insists that bringing in dynamic systems is crucial. The example, from my field, is that roboticists use dynamics (PDEs, signal processing, stats) instead of the embedded if-thens of production systems, due to both the "need for speed" in dynamics as well as flexibility. This has the additional benefit of suggesting a more robust model that includes both biology and algorithmic models, and Eliasmith invites both computational Neurology AND psych to converge in testing his new architecture. Emailer's Reference: Unified Theories of Cognition (William James Lectures).
DO take advantage of the author's/ Amazon's wonderful look inside feature, Dr. Chris even gives an extensive history in that look of what led up to this new model. One interesting gap I found in the otherwise extensive bib (nearly 30 pages) is Bach. You might find his book/dissertation interesting in summary and contrast to many of Chris' unique perspectives: Principles of Synthetic Intelligence PSI: An Architecture of Motivated Cognition (Oxford Series on Cognitive Models and Architectures). Both of course mention Newell extensively.
As a completely unscientific, but fun exercise, whenever I evaluate a title on cognitive architecture, I immediately go to the index and look for the seminal (at least in programming/robotics) word: "recursion." Newell and Chris don't index it specifically, Bach has half a dozen entries and hundreds cross referencing hierarchy and reflection, similar concepts, as do Chris and Newell. If you like or are curious about this concept, ubiquitous in this field but not always re-cognized, you'll greatly enjoy Corballis: The Recursive Mind: The Origins of Human Language, Thought, and Civilization. In fact, mathematically, there are recent ("matrix like") theories extending recursion to the mathematics beneath the entire universe, and from a physicist's frame, Dr. Tegmark is one of the first to explore it extensively: Our Mathematical Universe: My Quest for the Ultimate Nature of Reality.
Most recent customer reviews
Chris Eliasmith suspect that we don't just want to know how a single neuron works, or what brain centers...Read more