|
|
|
|
Before, we have just commented in the introduction that the firsts investigators had were created intelligent systems based in the model of McCulloch. In 50's is when is created a system with sucess, it called Perceptron of Rossenblatt. It was a visual system based in the recognition of patterns which could solve a lot of problems, but all the enthusiasm dissapared soon.
In 60's was when Alan
Newell and Herbert
Simon working in demostrating theorems and chess by PC created a
program called GPS (General
Problem Solver). This was a system that the user could define an emviroment
in function of an objects and operators that could be applicate over
there. This program could work with Towers of Hanoi and criptoarithmetic
and other similar problems, operating with a formalized microcosmos
that had the parameters inside of them they could solve problems. The
GPS could not solve problems
in a real world, and not to take importants decisions. The GPS
could operate with heuristics conditions(learn with owns discovers)
that gone to the objetive througth try and error. The first expert system was called Dendral, an interpreter of spectrogram created in 1967, but the with most influence was Mycin of 1974. The Mycin was capable to diagnotic problems in the blood and write a correct medication, it was a succes in this years and also was used in hospitals(like Puff, a variant of Mycin but with a common use in the Pacific Medican Center in San Francisco, USA). In 80's, was developed specials languages for to use with Artificial Intelligence like LISP or PROLOG. In this years was developed some betters experts systems, for example, the EURISKO. This program could to perfect its own set of heuristic rules automatically, by induction. |
|