Because of my background in physics, where I did my thesis work using a very
large MC simulation to study phase transitions, I have the expectation that MC
simulations should work much better when properly biased. In physics, we have a
theoretic basis for the bias, often from Boltzman or other appropriate
statistics. Unfortunately, not so in Go.
Here is a pointer to all of our work on computer go:
http://users.soe.ucsc.edu/~charlie/projects/SlugGo/
The papers to look at are the last 3 at that link, all from Jen Flynn. Her
project explored that question from a sampling perspective:
• Here are three unpublished quarterly reports from Jennifer Flynn.
• SlugGo Summer 07 Report: Predicting the Winner of 9x9
Computer Go Games using Patterns as Evaluators
• SlugGo Winter 08 Report: Using 5x5 Tiles with Libego
• SlugGo Spring 08 Report: Using Patterns in Libego Playouts
Unfortunately, we did not get anything we could call a positive result.
Cheers,
David
On 26, Apr 2010, at 2:53 PM, Hendrik Baier wrote:
> Hello list,
>
> I am a Master's student currently working on his thesis about certain aspects
> of Monte-Carlo Go. I would like to pose a question concerning the literature
> - I hope some of you can help me out!
>
> My problem is that I can't find many papers about learning of MC playout
> policies, in particular patterns. A lot of programs seem to be using Mogo's
> 3x3 patterns, which have been handcoded, or some variation thereof. A lot of
> people have tried some form of pattern learning, but mostly to directly
> predict expert moves it seems, not explicitly optimizing the patterns for
> their function in an MC playout policy. Actually, I am only aware of
> "Computing Elo Ratings of Move Patterns in the Game of Go", where patterns
> have been learned from pro moves, but then also successfully used in an MC
> playout policy; and "Monte Carlo Simulation Balancing".
>
> Considering the huge impact local patterns have had on the success of MC
> programs, I would have expected more attention towards automatically learning
> and weighting them specifically for MC playouts. There is no reason why
> patterns which are good for predicting experts should also be good for
> guaranteeing diverse, balanced playout distributions. Have I missed something?
>
> Or how did your program come to its patterns? I'd be interested. Did you
> maybe even try learning something else than patterns for your playout policy?
>
> cheers,
> Hendrik
> _______________________________________________
> 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
