DeepBlue

A highly parallel computer developed by IBM to play chess.

(See KasparovVsDeeperBlue ISBN:0713483229)

It uses essentially a brute force algorithm.

It produces moves as good or better than a grand master.

An interesting challenge is how to extend that power to make explanations for its choices of moves that are intelligible to a human. How does one post process such a deep-tree analysis into a summary that is concise and understandable, the way first class chess game annotations do?

• Nice expert system. But it doesn't learn, does it?

• Chess Analysis just tell you why that move is good. It doesn't mean that it is focused on winning necessarily. Just a Communal Acquired Knowledge, isn't it? You could attach a general comment to that line of code/tree choice. That would then make it the same thing then, wouldn't it? --StarPilot