Trending: With its Elasticsearch distribution, Amazon Web Services sends more shockwaves through open-source software
Go game
After winning its matches against European Go champion Fan Hui, the A.I. program AlphaGo is taking on South Korea’s Lee Sedol, a legendary figure in the game of Go. (Credit: Google DeepMind)

This month’s human-vs.-machine Go match between South Korean legend Lee Sedol and Google DeepMind’s AlphaGo AI program is a teachable moment – not only for experts in the field of artificial intelligence, but for aficionados of the millennia-old game of Go as well.

The five games in the $1 million challenge will be streamed live online from Seoul, with the first game due to begin at 8 p.m. PT Tuesday.

There’ll be online commentary, but if you’re looking for more of the human touch, show up at the Seattle Go Center, at 700 NE 45th St. in the University District. The center will be streaming each match on a big screen, and if you’re a newbie, you can learn how to play the game while Lee contemplates his moves.

“This is the ‘John Henry’ moment for the 21st century,” Brian Allen, manager of the Seattle Go Center, told GeekWire in an email. He’s referring to the 19th-century folk tale about a “steel-drivin’ man” who was pitted against a steam-powered hammer.

You could also call it a Garry Kasparov moment: Like IBM’s DeepBlue chess matches against Kasparov in the 1990s, AlphaGo’s game-playing skill gets into an area that was once thought the exclusive province of human intelligence. With 10170 legal board positions, Go is considered more complex than chess (which is said to have somewhere around 1050 legal board positions).

Google DeepMind says AlphaGo takes advantage of an innovative double-barreled strategy to accelerate the pace of its learning – a strategy that can be applied to a wide range of real-world tasks. AlphaGo’s five-for-five triumph over European champion Fan Hui last October marked the first time that a professional Go player was beaten by artificial intelligence in a fair fight.

Lee Sedol is regarded as a significantly better player than Fan Hui, but the crowdsourced odds on Good Judgment give the edge to AlphaGo.

So who does Allen think will win?

“I don’t know,” he told GeekWire. “The big question is, how much has AlphaGo improved since October? For a human, it would have to be a remarkable improvement, but in this case, you have an artificial intelligence program backed by Google’s network.”

The first game is likely to provide a good indication of how the five-game match will go. And it’s a good opportunity to start in on free 10 visits to the Seattle Go Center. Who knows? You may get hooked on the game, which would be a decisive win for Allen and his fellow Go enthusiasts.

Subscribe to GeekWire's Space & Science weekly newsletter

Comments

Job Listings on GeekWork

Web DeveloperRad Power Bikes
UX DesignerTINYpulse
Electrical EngineerRad Power Bikes
Sr SDETSkyKick
End User Support 1City of Bellevue
Find more jobs on GeekWork. Employers, post a job here.