The Computer and the Brain: Second Edition (Mrs. Hepsa Ely Silliman Memorial Lectures) Paperback – July 11, 2000

I could not locate my old copy of von Neumann's book so ordered a new one. The book, if a bit dated in its outlook, is still a fascinating intro to AI. What I can't figure out is why the Churchlands were asked to write a preface to the book, I wish they had at least done their homework or just checked their facts; mistaking William Church (a relative unknown in spite of a patent on typesetting machine dating back to early 1800's) for Alonzo Church, a towering figure in the foundations of computing is plain unacceptable.

Reza Langari

After 50 years, this book by the genius John von Neumann is still relevant in many aspects. I wish I had read this before I started my cognitive science education or before I have written my cog. sci. thesis. Neumann's insights into the architecture of the information processing of the brain is what many scientists today consider a nearly standard framework.

Anybody in interested in the intersection of computing science and brain research should read this short and sharp book, not only for its contents but also for Neumann's style.

Very good book

Today, the idea of the computer (and more recently, the network) as a model of the human brain, and by extension, the human mind, is not a particularly novel idea. But in von Neumann's time, it was in many ways a radical notion. Computers at that time were sometimes referred to as "mechanical brains," but that was more a metaphor for the masses than an idea that was taken seriously. In 1958, there were two types of computers, neither of which had anything approaching the memory or the speed or the complexity of the human brain. There were mechanical analog computers, like the Bush Differential Analyzer, that could be set up to provide numeric approximations of differential equations, and a few very large, expensive, and slow electronic computers that could be "programmed"- I think "configured" is a more descriptive term- to solve certain numeric computations, or approximations. It was more of a programmable calculator than a computer in the modern sense, as it was not able to change its own program code while running.

But von Neumann (and to be fair, a few other far sighted mathematicians like Turing) understood that the contemporary computers of his time were but a glimpse of what was theoretically possible, given a more advanced technology. It might even be possible, he thought, to achieve a level of complexity and speed that would approximate that of the human brain, and implicit in that was a very important insight: The specific hardware used to implement a given program was unimportant. While there were certainly parallels between the function of the brain and the electronic computers of the 1950s, what was important was not that both used electrical impulses, or that both had certain types of electrical switches. Here, he applies Turing's notion of a "code"- a set of instructions that allows any machine to perform in a manner identical to any other machine, given a suitable set of specifications and enough speed. And if the brain is , in fact, a deterministic biochemical machine, it should be possible to instantiate an equivalent of the human brain, and by extension, human mind, in artifical hardware.

There are of course several objections to that idea, like Hubert Dreyfus' objection regarding whether all knowledge can be formalized, and what I think of as the Wittgensteinian objections, that thought, like language, must be based on interaction with other actors in the world. But that's not to say that some kind of intelligence, some kind of self-aware computational being can be instantiated in some conceptual advanced hardware system. Regardless, von Neumann was among the first to realize that the true power of the computer lie not in faster and faster solution of computational problems, but in the abstract manipulation of symbols, and that through this manipulation, the computer of the future would be able to perform the actions of any machine, and possibly of living organisms as well. This is stimulating reading both for its historical import and for its logical presentation of issues that are still important in the cognitive sciences.

Von Neumann is one the brilliant mathematician and an expert of computer logic. This book is dated, manuscript written in 1957, but from the historical perspectives it still well worth reading. Neumann's pioneering work lead to considerable advances in computers and his ideas lead to advances in computer automation and robotics. The thoughts of Klara von Neumann, the wife of John Neumann provides a brief sketch of events that lead to the presentation at the Silliman Foundation lectures at Yale University. Neumann was diagnosed with bone cancer that confined him to the wheelchair. His health deteriorating by the day until his death in early 1957, unable to deliver the prestigious lecture and unable to complete the manuscript for the lecture, Yale University eventually published his partly-completed manuscript as a part of the prestigious Silliman lectures.

This book is described in two parts; the computer and the brain. The basic concepts of analog and digital procedures, the characteristics of digital machine types and their basic components, memory-stored controls, memory capacities, and the concept of access time are discussed with regards to the machine. In the second part the author discusses the structure and function of human brain and compares the common characteristics between the brain and computer. The author provides a comparative analysis of the nerve cell (neuron); how it generates and propagates nerve impulses compared with generation and propagation of computer messages. The author looks at the complexity on neurons and its functions; the nature of the nerve impulses, the process of its stimulation, digital character, the problem of memory within the nervous system. Although the author still refers to vacuum-tube machines, the flip-flops, and transistor technology, but the basic concepts underlying the development of memory elements in a computer is well worth the reading.

The recent advances in automation and robotics illustrate the early contribution of von Neumann in this field. In a recent study, Christopher Macleod and his colleagues in Aberdeen, UK, have created a robot that evolves like a living species in biological evolution. When the incremental evolutionary algorithm (lEA) realizes that its evolutions are no longer improving the robot's speed it freezes the neural network it has evolved, denying it the ability to evolve further. The sensors determine what it needs to carry out a given task most effectively. As animals evolved, the robots can evolve similarly. The robot can also adapt to newly acquired vision, and learn how to avoid or seek light when given a camera. This is just like the way the brain evolved building up in layers.

1. Minds, Brains, and Computers: An Historical Introduction to the Foundations of Cognitive Science (Blackwell Philosophy Anthologies)
2. Toward Brain-Computer Interfacing (Neural Information Processing)
3. From Computer to Brain
4. The Conscious Mind: Programming The Brain-Computer
5. Creating Brain-Like Intelligence: From Basic Principles to Complex Intelligent Systems (Lecture Notes in Computer Science / Lecture Notes in Artificial Intelligence)