Re: [computer-go] Global RAVE
Similar to normal RAVE: Initially it was weighted 1:0 with UCT, eventually 0:1 with UCT. I think it reached zero weight after about 5000 games. Am 25.07.2009 um 00:59 schrieb Peter Drake: Did you weigh these data as heavily as the raw Monte Carlo data, or discount them somehow? On Jul 24, 2009, at 3:50 PM, Isaac Deutsch wrote: My program performed worse with this (about 50 elo), but I didn't combine it with RAVE (just pure UCT with this). Am 25.07.2009 um 00:38 schrieb Peter Drake: Is anyone also using a global table of moves made after ALL nodes in the search, a sort of global RAVE table? It would be noisier still, but it would fill with useful data very quickly, no? Peter Drake http://www.lclark.edu/~drake/ ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/ Peter Drake http://www.lclark.edu/~drake/ ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/ ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Global RAVE
My program performed worse with this (about 50 elo), but I didn't combine it with RAVE (just pure UCT with this). Am 25.07.2009 um 00:38 schrieb Peter Drake: Is anyone also using a global table of moves made after ALL nodes in the search, a sort of global RAVE table? It would be noisier still, but it would fill with useful data very quickly, no? Peter Drake http://www.lclark.edu/~drake/ ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Global RAVE
Did you weigh these data as heavily as the raw Monte Carlo data, or discount them somehow? On Jul 24, 2009, at 3:50 PM, Isaac Deutsch wrote: My program performed worse with this (about 50 elo), but I didn't combine it with RAVE (just pure UCT with this). Am 25.07.2009 um 00:38 schrieb Peter Drake: Is anyone also using a global table of moves made after ALL nodes in the search, a sort of global RAVE table? It would be noisier still, but it would fill with useful data very quickly, no? Peter Drake http://www.lclark.edu/~drake/ ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/ Peter Drake http://www.lclark.edu/~drake/ ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/