Playing Go
Active from 1998 - 2001
Using the SANE and ESP methods, we have evolved networks to play go on small boards. When evolved against a computer program, the networks quickly learn to utilize weaknesses in the opponent, and exhibit aspects of general go-playing behavior. With fitness sharing, shared sampling, and hall-of-fame techniques, networks can be evolved against each other, and they reach a higher playing level that when evolved against a single opponent. Evolving specialists for different parts of the board also leads to improved performance.
Neuroevolution: Harnessing Creativity in AI Model Design Sebastian Risi, David Ha, Yujin Tang, Risto Miikkulainen To Appear In , Cambridge, MA, 2025. MIT Press. 2025