Date: Tue, 5 Aug 1997 15:26:18 +1000 From: mjb-AT-comserver.canberra.edu.au (Michael Booth) Subject: Re: inductive logic Greetings to the Technology group JF Koh <koh-AT-cleo.murdoch.edu.au> wote: >Thanks for the leads, folks. > >I did a double-take when someone told me that it has been done. Just had >to double-check and be sure. > >Would anyone be interested in a discussion about IBM's chess-playing >computer, Deep Blue? > > > > --- from list technology-AT-lists.village.virginia.edu --- There isn't a whole lot to discuss. Chess, being a game that is finite - there are a very large, but calculable set of possible moves, there is no great challenge to geting a computer to play, and presumably ensureeither a checkmate or a draw. The algorithm goes something like this, according to Daniel C. Dennett. Draw the decision tree of all possible chess games ( a vast but finite number). Go to the end node of each game, it will either a win for white, a win for black, or a draw. Color the node white, black or gray depending on the outcome. Work backwards one whole step (one white plus one black move) at a time; if on the previous move all the paths from any one of white's moves lead through all black's responses to a white colored node, color that node white, and then move back again. Do the same for any paths that guarantee a win for black. Color any other nodes gray. At the end of this procedure (way past the end of time), you will have identified all the games and all the moves that must result in a white win, leasding from the end game all the way back to any of the twenty legal opening moves. Now ply, sticking only to those moves that lead to a white win, and avoid any move that might open up the possibility of a black win. The worst you can do is draw. This is the theory, not remotely possible or feasible in practice, because the set of possibilities is enormous. Clearly those who programmed Deep Blue, Deeper Blue, whatever the machine is called have come up with algorithms that do not follow the above procedure, but algorithms that are a great deal more heuristic and less mechanical, at least to the point where they can beat the best that human beings can put up. Most things are, according to Dennett, capable of being the intentional or unitentional product of single or combinations of algorithms, whatever that tells about the respective capacities of human beings and machines constructed by human beings. Nice to get on-line, I hope this fits in with the tenor and tone of the spoon collective and technology. See ya round. - Take Care Michael Booth Associate Lecturer Faculty of Communication University of Canberra Ph: 61 6 201 2161 (w) 61 6 241 2591 (h) PO Box 1 Fax: 61 6 201 5119 Belconnen, ACT, 2616 Email: mjb-AT-comserver.canberra.edu.au AUSTRALIA ============================================= I Know It's Only Rock'n'Roll. But I like it, like it, like it, yes I do. ============================================= -----BEGIN PGP PUBLIC KEY BLOCK----- Version: 2.6.3i mQBPAjH0m1kAAAECAMNq6swTGMMKWIJwyalIhAKzYQzTq4hf/mUwsOczRySRlnDV d6/e4bUVD9+zKnomL50MhhQlDhxXLS8jd8XNyNEAEQEAAbQvIk1pY2hhZWwgQm9v dGggPG1qYkBjb21zZXJ2ZXIuY2FuYmVycmEuZWR1LmF1PiKJAFUCBRAx9JusLS8j d8XNyNEBAbwqAf9jM4qjpfgwhfUdUYmamVycStSAioKvp4hY88bMIVycWA9LaDQc Zh5HGY6adNRTmCpsMPuGnTVIBWN00mc0a38B =2eB+ -----END PGP PUBLIC KEY BLOCK----- --- from list technology-AT-lists.village.virginia.edu ---
Display software: ArchTracker © Malgosia Askanas, 2000-2005