It was just a phrase to demonstrate the difference between making a program work at playing chess and actually knowing chess so you can program it to make good evaluations.
Obviously it's not teaching in a human sense of the word.
It was just a phrase to demonstrate the difference between making a program work at playing chess and actually knowing chess so you can program it to make good evaluations.
Obviously it's not teaching in a human sense of the word.
Programmers have tried to create programs that learn that way; so far, the results have been a modest success, but why do that when you could just write a brute force algorithm and bypass the "learning" phase entirely? Pretty much all of the top 50 chess engines use brute force.
I understand what you are saying. I've spent some time thinking about what would constitute artificial intelligence and I believe that the ability to learn would be a requirement for that. That's why I only disagreed semantically, it's perfectly valid to say that they teach it.
However, in my own human mind I want to feel superior to machine because I actually "learned" to play chess. :-)
Yeah, which is really teaching it the chess.
I disagree with you semantically, Scottrf. My feeling is that we create an algoritm that determines the moves. To say that we are teaching it to play would entail learning on the part of the program.
If we were able to create a program that learned, theoretically we could give it the rules of chess and it would develop to grandmaster level by experience. Perhaps things have changed now, what I remember is that when a Kasparov threw something new or unexpected at Deep Blue the programmers were up all night changing the program for the next game.