Automated Go
The aim of this project is to investigate the application machine learning techniques to the ancient game of Go.

Go is an ancient asian board game of two players. Despite much research effort, computer Go has proven itself a difficult challenge, with the world's strongest Go program being unable to compete with weak amateur human players. The main reasons for this are that the game's huge branching factor defeats brute force search and it is difficult to write an efficient function to quantitavly evaluate Go positions.

 

 
 
 
The Game
Computer Go
Project Aims
Links
Publications
 
Home