Computer Go

Go-Playing Programs

Igowin

AlphaGo

Inspired new moves, such as the early 3-3 invasion. See https://senseis.xmp.net/?AIRevolution and AlphaGo Teach

Challenges:

• playing with handicap
• different board sizes
• multiple rules
• different komi
• maximizing the score

Open problems:

• pass-adversary, cyclic-adversary, see https://goattack.far.ai/
• requires special input for whether ladder works or not
• flying dagger joseki weakness, and possibly a few pincer joseki
• mirror Go wins as white
• train a net to “see” future states to be able to adjust its current policy
• playing like a human would at a given level

SAI

• Sensible Artificial Intelligence that plays with handicap and targets high scores in 9×9 Go

KataGo

• Opening books for 7x7 and 8x8

KaTrain

AhQ Go

BadukAI

Gridmaster

SmartGo One