Pasky’s Log

Today I held a presentation on the “complete” state of art in implementation of AI playing the board game of Go (much harder than chess programs), showing hopefully all the current widely applicable results, most successful strategies and worst current problems (in my interpretation, of course) – it is available on my Go Stuff page. Some neat algorithms and general tricks and not too much math can be found inside. ;-)

The presentation has a section introducing basic Go rules and tactics so it should be suitable even for AI researches not very familiar with Go (there is an appendix .zip archive containing example sequences for the rules, and also examples of some games and playout strategies). Of course it is not perfect and some parts assume accompanying explanations and few formulas on blackboard, but I think it could still be good material to give quick overview on currently used approaches with sufficient depth to satisfy a computer science scholar, plus quick references to the papers.

P.S.: One IMHO important thing I now realize I’ve missed in the “open problems” section is parallelization on GPU.