Tuesday, November 27, 2007

Language, Computers and Intelligent Systems

The successes of hardware technologies for Computers - transistors, integrated circuits and microprocessors - brought in a certain euphoria to Computer Scientists who set their sights pretty early on building the 'intelligent computer'. Spurred on by science fiction this 'Romantic Era' of 'Artificial Intelligence' made many pursue the holy grail of intelligence - human language processing.

Philosophers however were however interested in answering questions such as, the nature of intelligence, can a non-biological system really become intelligent, is intelligence a single faculty, is it inherited, how does learning occur, do creativity or intuition play any role in intelligence, and is self-awareness the ultimate test of intelligence.

Chief among these of course is the question whether intelligence is possible in artificial systems. The well known argument against its possibility is Searle's 'Chinese Room' argument, while counter arguments such as the 'Systems Reply' supported the view that intelligence was indeed possible in machines.

Two distinctly divergent schools pursuing the automation of intelligence defined their pursuit as follows:
  • the study of mental faculties through the use of computational models
  • a branch of Computer Science concerned with the automation of intelligent behaviour
The first group was more concerned with the study of human intelligence while the second was more interested in building intelligent computer systems.

A more pragmatics definition of artificial intelligence is that it is the 'study of how to make computers do things which at the moment humans do better'.

It appears that much of the confusion about the debate lies on the usage of terms such as intelligence, learning, knowing, understanding that have been up to now, associated almost solely with humans. The way forward is to do with the elevation of the meanings of these words to an abstract level so that there is a generic kind of intelligence for instance which has many specializations: Human intelligence being just one of them (and hence by definition different from machine - or indeed animal - intelligence.

No comments: