David Silver wrote: >> because the previous approaches were not optimized for such a small >> boards. > >I'm not sure what you mean here? The supervised learning and >reinforcement learning approaches that we compared against are both >trained on the small boards, i.e. the pattern weights are specifically >optimised for that size of board.
Ah, sorry. I missed it. >I agree that the handcrafted policy from Fuego was not optimised for >small boards, which is why it performed poorly. But perhaps this is >also interesting, i.e. it suggests that a handcrafted policy for 9x9 >Go may be significantly suboptimal when used in 19x19 Go. So >automatically learning a simulation policy may ultimately prove to be >very beneficial. It is surprising Fuego was far worse than uniform random on 5x5 and 6x6. These results show the particularity of the very small boards. I suppose 5x5 is very different from 9x9 but 19x19 is not so from 9x9. >> I'm looking forward to your results on larger boards. > >Me too :-) >Coming soon, will let you know how it goes. >But I hope that others will try these ideas too, it's always much >better to compare multiple implementations of the same algorithm. Could you give us the source code which you used? Your algorithm is too complicated, so it would be very helpful if possible. -- Yamato _______________________________________________ computer-go mailing list [email protected] http://www.computer-go.org/mailman/listinfo/computer-go/
