|
|
There was no clear definition for AI. The most widely accepted was Marvin Minsky's definition
"artificial intelligence is the science of making machines do things that would require intelligence if done by men"
In general, three distinct definitions of AI resulting from different motivations could be identified. The first definition was that AI was a type of advanced computer technology. The second one was that AI was simulating human behavior and cognitive processes on a computer. The third one was that AI was the study of nature of the whole space of intelligent minds. Each of these definitions represented one of the three established areas of: computer science, psychology and philosophy. However, early work in AI had basically been of an exploratory nature. The initial concern of AI researchers was to find possible ways of doing intelligent things. Therefore, early AI work, however much it was claimed otherwise, had addressed a different field from that of psychology. Psychology was concerned with actual way people do things; AI had been attempting to map out the possible ways that those things could be performed on a computer
Figure 1: Three distinct representations of AI
AI had been a source of theories for psychologists. This attention affected AI towards a change of emphasis in favor of cognitive simulation. At the same time, the question of whether machine could think arise. Alan Turing defined the question as follows
'I propose to consider the question, "can machine think?" ... I shall replace the question by another ... The new form of the problem can be described in terms of a game which we call "imitation game". It is played with three people, a man (A), a woman (B), and an interrogator (C) who may be of either sex. The object of the game for the interrogator is to determine which of the other two is the man and which is the woman
The question was 'What will happen when a machine takes the part of A in this game? Will the interrogator decide wrongly as often when the game is played like this, as he does when the game is played between man and woman?' These questions effectively replace the original question, 'can machines think?'. The Turing's Test was the experiment which a machine took the part of A in the above game. If an interrogator could not tell whether he or she was communicating with a human being or a machine, then the machine could be considered intelligent
The initial phase of AI research led to a general realization that world knowledge lies behind human ability and intelligent behavior. AI, in its attempt to produce intelligent behavior on a computer, was limited by the amount of knowledge about people's general knowledge. Other disciplines, such as psychology, had been unable to provide AI with an explicit exposition of the general knowledge that people showed in their behavior. Thus, some attention was focused on intellectual behavior. This was the kind of knowledge which a professional expert has. The attempt to deal with only a small domain of human knowledge reduced the complexity of the task faced by an AI researcher and enabled him to offer a working system to users on a reasonable time scale