Bob Stark: Expert Move Prediction in Go

With the guidance of David Helmbold, Bob Stark implemented machine learning methods in R Statistical Environment to further improve the skill of SlugGo, the Go-playing program developed at UC Santa Cruz. The game of Go is much harder for an artificial intelligence to play than most other games. A computer program has beaten the best Chess player in the world. The best Go program is still rather horrible compared to experts in Go, however. The American Association for Artificial Intelligence (AAAI) has therefore named Go a AI Grand Challenge. One of the features of SlugGo that is different than some other competitors is that it utilizes the Go proverb, "your opponent's move is your own"; it considers the values of close moves to the opponent when deciding whether to make a particular move. Bob's particular project was to discover the best way to go about considering these values. Future work, handed off to Dr. Helmbold, will be to further optimize the machine learning methods to provide the SlugGo team with a deliverable to add to the playing engine itself.