. Aretha Franklin wanted r-e-s-p-e-c-t, but frankly, my dear, i don't give a damn. this note is not a riposte to the latest epistle in a stream of libelous personal abuse, but a clarification intended to stimulate thought among any young-at-hearts that may also subscribe to this forum.
why would anyone want to program a computer to play an old board game? different people have different reasons because they have different motivations. for some, most notably Arthur Samuel, the attraction is self-evidently the kudos of fathering an electronic alter-ego that has a chance of achieving fame and glory by becoming the alpha male through gladiatorial olympic combat on the world's stage; the same motivation that led IBM to spend quite a lot of advertising budget to beat Gary Kasparov. for others like Jacques Pitrat, it is mathematical inquisitiveness. i mention these two because Pitrat overrode Samuel's veto of my IJCAI 79 paper on the grounds that i hadn't proved my fledgeling ideas by writing a program like he had done in 1952 (which did little more than demonstrate that Johnnie von N's 1928 minimax theorem was right in the first place). Thanks, Jacques! btw, i didn't find out about my debt of gratitude to Pitrat until decades later when Feigenbaum's old archives were digitised and put on the web. and for yet others, myself included, computer Go is a convenient experimental testbed for scientific enquiry into the nature of intelligence, following a path trodden by pioneers like John von Neumann, Alan Turing and Herbert Simon. Go is a convenient testbed for theoretical AI research because it is a narrowly-scoped information domain, one free from the foggy hairyness of the physical world, which thereby enables one to ponder on experimental models of knowledge representation and processing without becoming bogged down by also having to distinguish signal from noise, a significant hurdle that real-world robots have to confront.
_______________________________________________ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
