Computers and Thought

Rating:
Summary: Waiting to Think
Review: It looks like Alan Mathison Turing's suggestion stands the test of time: come up with a program that apes a child's mind and then put the computer through school. That may be what it takes to bring COMPUTERS AND THOUGHT together in the first real thinking machine. Proving theorems is good training in problem solving, which we learn along with verbal training, concept formulation, decision making under uncertainty, and social behavior. The first high order artificial intelligence, the logic theorist, finds proofs for theorems but gets into trouble when answers are beyond rote recalling of the table of integrals. We pull in intuition, transformations and tricks beyond elementary calculus, but the computer problem solves according to the way it was designed. So if computers could be designed to learn from routines, they would recognize patterns in proofsearch procedure and then strategize to repeat the patterns in later proofs. At the last page I felt energized from having gotten through such a challenging, excellently organized, and fascinating book. Editors Edward A Feigenbaum and Julian Feldman make a major contribution to information on artificial intelligence. So their book fits in with George Dyson's DARWIN AMONG THE MACHINES, Stan Franklin's ARTIFICIAL MINDS, David Freedman's BRAINMAKERS, John Haugeland's ARTIFICIAL INTELLIGENCE, or Clifford A Pickover's COMPUTERS AND THE IMAGINATION.