Lucas, Simon M
sml at essex.ac.uk
Mon Jan 20 05:22:35 PST 2014
I think the move prediction methods are also relevant e.g.
see Wistuba et al, CIG 2012:
From: computer-go-bounces at dvandva.org [mailto:computer-go-bounces at dvandva.org] On Behalf Of Cameron Browne
Sent: 20 January 2014 13:14
To: computer-go at dvandva.org
Subject: [Computer-go] RAVE-Tiling?
> Date: Mon, 20 Jan 2014 12:07:11 +0100
> From: Stefan Kaitschick <stefan.kaitschick at hamburg.de>
> To: computer-go at dvandva.org
> Subject: [Computer-go] RAVE-Tiling?
> Has anybody ever experimented with tiling the board with a set of possibly overlaping smaller tiles, and updating the RAVE statistic for those tiles, whenever the same pattern comes up for that tile in the tree? This would probably be more reliable than AMAF, and fill faster than RAVE, so it might be a useful addition. There could even be multiple sets of different sizes. I'm proposing fixed borders for the tiles. Another approach might be to introduce dynamic tiles in areas that have been found by some measure to be critical.
The papers on Contextual Monte Carlo might be relevant:
A. Rimmel and F. Teytaud, "Multiple Overlapping Tiles for Contextual Monte Carlo Tree Search," in Proc. Applicat. Evol.
Comput. 1, LNCS 6624, Torino. Italy, 2010, pp. 201-210.
J.-B. Hoock, C.-S. Lee, A. Rimmel, F. Teytaud, O. Teytaud, and M.-H. Wang, "Intelligent Agents for the Game of Go," IEEE Comput. Intell. Mag., vol. 5, no. 4, pp. 28-42, 2010.
Computer-go mailing list
Computer-go at dvandva.org
More information about the Computer-go