[Computer-go] RAVE Implementation Clarification

[Petr Baudis]

  Hi!

On Fri, Jan 21, 2011 at 07:51:06PM -0500, Kenny Lam wrote:
> According to the paper (or atleast as I understand it), the AMAF variance
> term for RAVE is calculated as:
> 
> sqrt(log(M) / m)
> Where m is the number of times that x was given an AMAF "virtual" update and
> M is the sum of m's for x and all of its siblings.
> 
> However, I've also downloaded the source for the TesujiRef engine from the
> plug-and-go's svn repository and it seems RAVE is implemented differently
> here.  In particular, I've noticed that the AMAF variance term is calculated
> as:
> 
> sqrt(log(N) / m)
> Where N is the number of real updated for x_parent and m is the number of
> virtual updates for x (in consistence with the terms used throughout this
> email).
> 
> Tesuji's AMAF variance term seems to only go up when x isn't chosen for a
> real update as opposed to a virtual update.  Which implementation is best?
> Am I simply misunderstanding one of the above?

  The former makes more sense, but note that many people do not use any
variance term for RAVE at all. Try out all three possibilities. :-)

> I've also noticed that the beta value used for RAVE is calculated
> differently.  The beta in Gelly and Silver's paper seems to decrease at an
> inverse sqrt rate where as the beta in TesujiRefBot decreases
> logarithmically.  Is the implementation in Tesuji more "modern" than the
> once described in Gelly and Silver's paper?  Thanks a lot for your help.

  You did not paste the exact formula used in TesujiRefBot, but there
are several variants of beta formulas out there. Few weeks ago, several
program authors described their exact node evaluation terms which should
give you a good overview of the state of art. None is probably strictly
better than the other, you will have to experiment and tune a lot.

-- 
				Petr "Pasky" Baudis
Computer science education cannot make an expert programmer any more
than studying brushes and pigment can make an expert painter. --esr