Cognitive Dynamic Systems: Perception-action Cycle, Radar and Radio

Neuroscience sometimes seems to suffer from ADD. It hops from computational models to brain rhythms and gamma inhibitory networks, to emotion, to rewards, to basal somatics, to... As it does so, it oscillates between discrete and time based explanations, often going so deep in one that it blinds itself to the others as also having pieces of the puzzle. Books like this, filled with innovative ideas, inspire and encourage unity and focus amidst those fractals. It takes a "radio guy" to see how "tuning" is a good word for phase transitions, whether we're tuning temperature in ice vs. water, or radio waves in MRI.Haykin is delightful, and has a "wonder" that makes it seem like Dynamical Systems was "just now" coming of age in cognitive science. Of course McCullough & Pitts (1943), Hebb (1949), Hopfield (81), Hertz (91) etc. have applied differential equations to time series in neuronal studies from early stimuli-response continuous distributions to collective computation, memory, and like Haykin, the perception-action cycle, for many decades. In fact, folks like Kurzweil (How to Create a Mind: The Secret of Human Thought Revealed) have moved away from DST to more probabalistic models based on a wider as well as more specific view of pattern recognition. See the reviews for Kurzweil's new book above for some additional valuable bib items in this area, in contrast to DST.By bringing a signal processing background to cognitive dynamics, Haykin is literally reinventing a field that, frankly, did great in neural networks, then stuttered in explaining cognition. Instead of getting all forest vs. trees, Haykin stays on solid comparative ground, using numerous examples from radar, radio, communications-- basically signal processing-- to reinvigorate the DST model of cognition. His work is breathtaking, sound and innovative.This book is highly technical, and won't ever garnish the "populist" audience of a Kurzweil, Shermer, Horstman, Thagard, etc. as it gets into detailed descriptions of how to create a stage by stage, time backward, cost-to-go function as the foundation for the dynamic programming algorithm, with attendant optimality equations. However, if you have any interest in the REAL applications of DST to, say, robotics, this volume is a must read."Give me a 555 and an op amp, and I can create anything" -- a joke about how important feedback and feed-forward are in correlative learning, but not lost on those of us doing practical work in robotic kinematics and vision. In an amazing synchronicity, both Kurzweil's model and Haykin have both recently contributed to advances in control-feedback between arm movement and neural network organization in my own engineering field.After all, discrete/continous blends in nature as we observe particles that are "really" waves, and an all-analog physical world that dances seamlessly with digital transformations. It is fascinating how innovatively Haykin uses memory, for example, as the "stabilizing" element in feedback loops for cognitive radar. Every sentence in this book has a new, out of the box, innovative way of looking at what I'd call phase transitions-- the movements BETWEEN feedback events. In those models, we need both continuous and discrete analytical tools.Just when you might have thought CDS was a fad for those loving to show off their PDE muscles-- Haykin has given it a new impetus and momentum, without shorting the differential equations, but putting it right back on the practical track of feedback and feedforward transmission, reception and control. Highly recommended for anyone wanting a very recent view of where DST is going, as well as specialists in robotics, cognition, signal processing, and many other fields. This book is groundbreaking, and Haykin's work will be seen as seminal in coming decades. DST held promise in cognition, that promise took a break, and Haykin, in this book, has revived that promise-- it's that good.Library Picks reviews only for the benefit of Amazon shoppers and has nothing to do with Amazon, the authors, manufacturers or publishers of the items we review. We always buy the items we review for the sake of objectivity, and although we search for gems, are not shy about trashing an item if it's a waste of time or money for Amazon shoppers. If the reviewer identifies herself, her job or her field, it is only as a point of reference to help you gauge the background and any biases.