[Pachi] Better semeai handling ?

lemonsqueeze lemonsqueeze at free.fr
Thu Apr 28 22:04:44 CEST 2016


Oh wow, yes it's definitely worth trying !

Do you have time for this ? I can try to look into it but might take
me a while (don't really know what i'm getting into)

I'm cleaning up my branch for the predictions, it's a bit messy right 
now. And i just noticed my moggy engine is pretty much a rewrite of 
replay/replay.c =)
How about grouping the alternative engines in one directory, that'd make 
it easier to find. Say other_engines/moggy.c or something like that ?


On 04/27/2016 11:32 AM, Petr Baudis wrote:
>    Remi trains the patterns as teams of features.  We just do frequentist
> statistics on patterns as a whole.
>
>    There should be some pretty rusty code that can feed patterns to
> Remi's MM tools and train the teams of features, which was part of the
> effort for the probdist playout policy.  I didn't bother trying this
> approach for the pattern priors.  Based on what you noticed (I never
> measured prediction rate - if you have some scripts/code for that, it
> could be nice to have merged!), it could be a pretty nice boost!
>
> On Tue, Apr 26, 2016 at 10:14:28PM +0200, lemonsqueeze wrote:
>> Wait, something's wrong.
>>
>> Shouldn't large patterns get around 34% prediction rate as in Remi's paper ?
>>
>>
>> On 04/26/2016 15:13 PM, lemonsqueeze wrote:
>>> Thanks, this makes more sense now.
>>>
>>> I guess i should apologize for my previous comment on moggy:
>>> I just ran some tests on kgs 6d games to get an idea of moggy's
>>> prediction rate and it does very well actually: 24%
>>> (i made the engine return the most frequent move from a number of
>>> moggy runs)
>>>
>>> It even beats the large pattern predictions which i thought would be
>>> better (around 21%)
>>>
>>> Does it sound possible that alphago's fast policy could have same
>>> prediction power and yet be significantly stronger wrt semeais ?
>>> Seems even more puzzling than before !
>>
>


More information about the Pachi mailing list