Artificial Intelligence : Usfssg Computers to Think about Thinking . Part 1

In 1950, Alan M. Turing, the late deputy director of the University of Manchester’s Computing Laboratory in England, proposed a novel test to determine whether a machine was capable of thinking. In thk test, an interrogator has a teletype conversation with a man and a woman, both of whom must try to convince the interrogator that they are the woman. At some point unknown to the interrogator, the man is replaced by a machine. If the interrogator is fooled as often by the machine as by the man, that machine can be said to have displayed intelligent behavior. 1 Some 30 years after Turing proposed this test, many aspects of human behavior have been simulated by a computer. Programs have been designed to play checkersz and chess,J prove mathematical theorems ,4,5 and even mimic the behavior of a paranoid human being.b Despite the success of these and many other programs, none of the researchers investigating what’s been variously cafled “applied epistemology” or “artificial intelligence” (AI) would claim this means the “thinking machine” has arrived. Instead, they would agree that these programs have contributed important information about human behavior, and how computers can simulate it. The first part of this two-part essay will review some of the theones AI researchers have developed to explain human “information processing.” The second part of the essay will cover some applications of AI research. These include programs used in robotics, programs that communicate with computer users in natural languages such as English, and “expert systems” which help chemists, physicians, and others perform decision-making tasks. The “pioneer” expert system, DENDRAL, will be discussed in some detail.T.~ AI grew out of the convergence of ideas in several different fields, and the availability of new technologies. According to Avrom Barr and Edward A. Feigenbaum, Stanford University, California, the single most important factor contributing to the birth of the field was the invention of the computer.9 They point out that human beings have always drawn analogies between mechanical devices and their own behavior. Computers, with their memories and information-processing abilities, naturafly invited analogies with the human brain. Shortly after digital computers became available, computer scientists began creating programs that, they hoped, would perform tasks generally considered to require intelligence. Their earliest efforts were directed at programming computers to solve puzzles, play games such as chess, backgammon, and checkers, solve mathematical theorems, and translate text from one language to another. The early computer programs performed these tasks, but not very well. For example, chess programs were successful at following the step-by-step instructions for moving chessmen. But computers couldn’t independently gen-