Human Anatomy - Nervous System, Neuroscience, Artificial Intelligence - General, Neurology, Neuroanatomy

Brainmakers

Name: Brainmakers
Author: David H. Freedman
ISBN: 9780671760793

by David H. Freedman

Write a review

Overview

The idea of building intelligent machines has fascinated scientists and lay people for centuries. But progress has been much slower than everyone predicted thirty-five years ago when artificial intelligence and robotics first emerged. Although computers can crunch numbers at staggering speeds, so-called "thinking machines" have achieved very little ability - outside of highly restricted domains such as chess - to reason in useful or interesting ways, or even to do simple things that children can manage easily, such as recognizing faces or moving through a room without bumping into furniture. But now a new approach to artificial intelligence promises to break that logjam. Brainmakers is an account of the revolutionary research taking place around the world, aimed at creating devices that are more like living brains than computers. David H. Freedman introduces us to the adventurous researchers leading this new field, as well as to their creations, including a robot farm where robots will be "bred" for intelligence; a jarful of chemicals that can recognize patterns, and act like a primitive life-form; a machine that experiences human-like brainwaves and mental disorders; and a new form of computer chip fashioned out of living brain cells. In colorful detail David H. Freedman shows how this radically new research has moved into a realm that transcends computer science, combining neuroscience, microbiology, evolutionary biology, and zoology. Brainmakers is the first account of the latest developments in this exciting area of science and technology.

In exciting and accessible detail, this book shows how groundbreaking research into the innermost workings of the human brain has moved into a new realm that transcends computer science. Modelled after natural rather than artifical intelligence, thinking machines may soon develop powers that rival--or exceed--those of the human brain.

Reviews

There are no reviews yet. Log in to write one.

Editorials

Publishers Weekly - Publisher's Weekly

Freelance science writer Freedman's compelling state-of-the-art report on the quest to build human-like thinking machines explores how the field of artificial intelligence is being reinvigorated through AI researchers' interface with neuroscience, biology and robotics. At MIT, Attila, a six-legged robot, crawls around, learning new skills by interacting with its environment. In Japan, scientists are making movies of the neuron-to-neuron flow of signals inside the brains of live rats; their ultimate goal is a ``wiring diagram'' illustrating how the human brain works. At UCLA, computer scientists have designed a ``robot farm'' where robots will ``mate'' by merging their programs; an occasional mutation will be added to imitate biological evolution. Even more science fiction-like are biophysicists' and AI experts' efforts to harness the self-organizing and memory capabilities of biomolecules (e.g., bacterial protein or RNA) which may one day replace transistors on microchips or even serve as the basis for ``biomolecular computers.'' (Apr.)

Jon Kartman

If you think supercomputers like Hal 9000, star of "2001", are inevitable, Freedman may surprise you. For after more than 30 years of trying to create an electronic brain that matches our glob of gray matter, he discloses, scientists and engineers readily admit they are still far from any such goal. Lest you start thinking the artificial intelligence field is totally barren, Freedman shows what's presently in it--computers ranging from some that can think and do logic but not much else to others that can almost sing and dance but don't have the sense to come in out of the rain. The scientific movers and shakers have come to a dead end in trying to make computer brains that look and act like R2D2 and instead are looking at how the brain evolved. Ironically, the lowly cockroach and other bugs may provide the inspiration for actually building a brain. Some of the newest developments are downright startling, allowing Freedman to assert that the next decade or so may see "thinking" computers that may make Hal 9000 seem outmoded in the real year 2001.