The Future of Go

Well thanks for making it this far! I'll finish with a quick look at the techniques research is focusing on at the moment.

The Future

Combinatorial Game Theory

Combinatorial game theroy involves segmenting the board into subgames, reasoning about the subgames independently, suggesting moves for each subgame, and then choosing the best move overall. Unfortunately this only works well for the end-game, as it is only then that subgames may be reasoned about independently.

Hierarchical task network planning

Another area of research is looking at playing Go using a goal driven approach, as opposed to the data driven ones we've looked at so far. Although AI literature deals with HTN planning quite extensively, using multiple agents that can compete against each other allows for adversarial planning and this looks like an interesting approach to learning Go.

Super computers

Perhaps there will eventually be enough computing power to brute-force search the game tree of Go, but will there be a good enough evaluation function? I hope not because Go is providing an excellent domain for AI researchers and i believe it will continue to bring out new ideas and approaches to game AI.

Go Go Go!

Hopefully I've conviced you of the size of the problem of playing Computer Go, and have interested you enough to look into it futher. The following are good places to start:

Perhaps I should mention in closing that there is a $2,000,000 prize for writing the first computer program to beat a Go master, so get Going

Remember you can visit the Message Store to discuss this essay. Comments are always welcome!. There are already replies in the thread, why not join in?