It's pretty easy to find cases where the AI blunders. Automatically finding these is not a bottleneck to improving performance. The bottleneck is programmer time to analyze and fix the problems, and test resources to verify that the fix did not make the program weaker in other situations. I have dozens of similar test cases for Many Faces waiting my attention.
David From: [email protected] [mailto:[email protected]] On Behalf Of terry mcintyre Sent: Friday, November 26, 2010 9:17 AM To: [email protected] Subject: Re: [Computer-go] Dave Devos vs Fuego Martin, how close were the 2nd and 3rd choices? Looking at the bigger picture: is it yet possible for programs to do automatic post-game reviews of losing games - possibly spending a few hours or days on a single game - and locate weak spots in their own behavior? Another question: at what point did Fuego realize that it had blundered? ( when did the "winning percentage" show a sharp decline? ) Did it "know" that it was ahead prior to that weak move which arguably cost it the game? How close do such estimates correlate with the estimate of strong players? When they differ significantly, is it possible to say whether the program or the players are more correct? Terry McIntyre <[email protected]> Unix/Linux Systems Administration Taking time to do it right saves having to do it twice. _____ From: "[email protected]" <[email protected]> To: [email protected] Sent: Fri, November 26, 2010 11:41:16 AM Subject: Re: [Computer-go] Dave Devos vs Fuego Ok, so it's probably specific to Fuego like David Fotland and Magnus said. Glad I could help ;) Dave _____ Van: [email protected] namens Martin Mueller Verzonden: vr 26-11-2010 16:49 Aan: [email protected] Onderwerp: Re: [Computer-go] Dave Devos vs Fuego Dave, I ran current Fuego on my laptop for 600K simulations. Similar to the other programs, it likes simply C7 with a .82 evaluation. However, the bad moves M19 and N19 are second and third in number of simulations. So there seems to be some systematic problem with the playouts and/or the tree search here, which makes this white group die often after those moves. Strange since N19 is not even a threat. Thank you for the test case. Martin _______________________________________________ Computer-go mailing list [email protected] http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
_______________________________________________ Computer-go mailing list [email protected] http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
