I was thinking the same thing. You can easily equip the value network with several outputs, corresponding to several settings of komi, then train as usual.
The issue with Japanese rules is easily solved by refusing to play under ridiculous rules. Yes, I do have strong opinions. :) Álvaro. On Tue, Mar 21, 2017 at 7:36 PM, <cazen...@ai.univ-paris8.fr> wrote: > > Why can't you reuse the same self played games but score them with a > different komi value ? The policy network does not use the komi to choose > its moves so it should make no difference. > > > > On 21/03/2017 21:08, David Ongaro wrote: > >>> But how would you fix it? Isn't that you'd need to retrain your value > >>> network from the scratch? > >> > >> I would think so as well. But I some months ago I already made a > >> proposal in this list to mitigate that problem: instead of training a > >> different value network for each Komi, add a “Komi adjustment†value > >> as > >> input during the training phase. That should be much more effective, > >> since the “win/lost†evaluation shouldn’t change for many (most?) > >> positions for small adjustments but the resulting value network (when > >> trained for different Komi adjustments) has a much greater range of > >> applicability. > > > > The problem is not the training of the network itself (~2-4 weeks of > > letting a program someone else wrote run in the background, easiest > > thing ever in computer go), or whether you use a komi input or a > > separate network, the problem is getting data for the different komi > > values. > > > > Note that if getting data is not a problem, then a separate network > > would perform better than your proposal. > > > > -- > > GCP > > _______________________________________________ > > Computer-go mailing list > > Computer-go@computer-go.org > > http://computer-go.org/mailman/listinfo/computer-go > > > > > _______________________________________________ > Computer-go mailing list > Computer-go@computer-go.org > http://computer-go.org/mailman/listinfo/computer-go >
_______________________________________________ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
