Vision by man and machine.

How does an animal see? How might a computer do it? A study of stereo vision guides research on both these questions. Brain science suggests computer programs; the computer suggests what to look for in the brain. T he development of computers of increasing power and sophistication often stimulates comparisons between them and the human brain, and these comparisons are becoming more earnest as computers are applied more and more to tasks formerly associated with essentially human activities and capabilities. Indeed, it is widely expected that a coming generation of computers and robots will have sensory, motor and even "intellec-tual" skills closely resembling our own. How might such machines be designed? Can our rapidly growing knowledge of the human brain be a guide? And at the same time can our advances in "artificial intelligence" help us to understand the brain? At the level of their hardware (the brain's or a computer's) the differences are great. The neurons, or nerve cells, in a brain are small, delicate structures bound by a complex membrane and closely packed in a medium of supporting cells that control a complexand probably quite variable chemical environment. They are very unlike the wires and etched crystals of semiconducting materials on which computers are based. In the organization of the hardware the differences also are great. The connections between neurons are very numerous (anyone neuron may receive many thousands of inputs) and are distributed in three dimensions. In a computer the wires linking circuit components are limited by present-day solid-state technology to a relatively small number arranged more or less two-dimensionally. In the transmission of signals the differences again are great. The binary (on-off) electric pulses of the computer are mirrored to some extent in the all-or-nothing signal conducted along nerve fibers, but in addition the brain employs graded electrical signals, chemical messenger substances and the transport of ions. In temporal organization the differences are immense. Computers process infonna-tion serially (one step at a time) but at a very fast rate. The time course of their operation is governed by a computer-wide clock. What is known of the brain suggests that it functions much slower but that it analyzes information along millions of channels concurrently without need of clock-driven operation. How, then, are brains and computers alike? Clearly there must be a level at which any two mechanisms can be compared. One can compare the tasks they do. …