Description: The first book to take VLSI into the analog domain and apply it to biology. It provides solid tools for research in artificial intelligence and neurobiology while illustrating powerful new applications for analog systems.
Customer Reviews
Review Summary: make silicon eyes and ears
Date: 2007-03-07
Details: The book describes an interesting niche in VLSI design. Most VLSI chips implement digital logic. But Mead took a different tack, emphasising the analog mode of operation of the transistors. In most digital electronics texts, this regime of current-voltage performance is mostly cursorily dealt with.
What Mead did was use this often where the current through the source and drain was some exponential function of the voltage at the transistor gate. An oversimplification, perhaps, but it captures the essence of the book. By tying together transistors, Mead was able to build circuits that emulated the performance of the eye and ear. The text then uses these to make silicon chips that might mimic the biological sensors.
The book also embodies Mead's approach to understanding the brain and its neural networks. He claims that the problem is very hard. And that we can usefully make progress by looking at the brain's input sensors. As these are much simpler to understand and implement.
Mead carried the ideas here into Synaptics. A Silicon Valley startup that he co-founded.
Sadly, the book is out of print. (Why??) The prices of $129 and higher by third party sellers are way excessive.
Review Summary: the best book about biologically inspired models
Date: 2002-08-15
Details: This is definitely a must-read book for researchers and students in the field of neuromorphic engineering such that they could learn how to engineer the biological systems...
Review Summary: The classic text for this kind of circuitry.
Date: 1997-06-07
Details: Assuming you know something about CMOS and VLSI
design, this is the classic text to cover the
broad base to get started in understanding how
one goes about designing actual hardware for
various neural network architectures. Both
analog and digital approaches are discussed, and
the circuits are clearly explained with lots of
schematics and plenty of derivative mathematics
that show why a particular approach has utility
for a given problem. There are a lot of new
books (Mead has a new one out) but they owe a
large debt to this book.
out of 3 Reviews
