University of Cambridge > Talks.cam > Inference Group > Bayesian Pattern Ranking for Move Prediction in the Game of Go

Bayesian Pattern Ranking for Move Prediction in the Game of Go

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Phil Cowans.

I will describe a system for predicting the moves of expert Go players by pattern matching. A pattern is defined as an exact arrangement of stones on the Go board associated with a hypothetical move.

The system learns to predict the moves of experts in a two stage process. Firstly millions of these patterns are automatically harvested from hundreds of thousands of records of Go games. Secondly these same games are used to learn an urgency value for each of the patterns to enable us to use them to select moves.

We have achieved the best currently published results for move prediction in Go, perfectly predicting 33% of expert moves and ranking 80% of expert moves in the top 10.

This talk is part of the Inference Group series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

© 2006-2020 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity