Collective computation in neuronlike circuits.
暂无分享,去创建一个
M odern digital computers are latecomers to the world of computation. Biological com puters-the brain and nervous sys tem of animals and human beings have existed for millions of years, and they are marvelously effective in processing sensory information and controlling the interactions of ani mals with their environment. Tasks such as reaching for a sandwich, recognizing a face or remembering things associated with the taste of madeleines are computations just as much as multiplication and running video games are. The fact that biological computa tion is so effective suggests that it may be possible to attain similar ca pabilities in artificial devices based on the design principles of neural systems. We have studied a number of "neural network" electronic cir cuits that can carry out significant computations. Such simple models have only a metaphorical resem blance to nature's computers, but they offer an elegant, different way of thinking about machine computa tion, which is inspiring new micro electronic chip and computer de signs. They may also provide fresh insights into the biological systems. Current research on this subject builds on a long history of efforts to capture the principles of biological computation in mathematical mod els. The effort began with the pio neering investigations of neurons as logical devices by Warren S. McCul loch and Walter H. Pitts in 1943. In the 1960's Frank Rosenblatt of Cornell University and Bernard Widrow, who is now at Stanford University, creat ed "adaptive neurons" and simple networks that learn. Widrow's Ada line (short for adaptive linear ele-