Daniel Bump of Stanford University reviews in this lecture the history of efforts to develop programs that play the ancient Asian game of Go. Such programs have been more difficult than those for chess, due to the greater complexity of pattern recognition and matching required.

Bump’s lecture covers the technology, early programming efforts, people, and politics of the game. It assumes background knowledge of Go, so it is more appropriate for players and for those interested in computer versions of Go and other strategy games.

The video presents Bump lecturing, and pans to his presentation slides; viewers are advised to halt the video at each slide since they are a bit difficult to read during the lecture. Bump also reviews several games of Google’s AlphaGo against human Go experts, notably Lee Sedol, the Korean Grandmaster, including Sedol’s only victory against the machine.

Modern Go programs have advanced from early hand-coded pattern matching of good plays to neural net and Monte Carlo methods. Bump explains these only briefly, and viewers would need some additional background in such technologies to benefit from the presentation.

Viewers interested in following the game itself, and finding opponents, should investigate one of the current popular Internet Go servers at http://pandanet-igs.com. Those interested in Google’s AlphaGo and related artificial intelligence (AI) methods will find much to discover at https://deepmind.com/.