Google DeepMind’s AlphaGo was defeated Sunday by Lee Se-dol, the South Korean player that the program had defeated in three consecutive games.
Advertisement
In game 4 of a best-of-five match between Google’s AlphaGo program and master Go player Lee Sedol, Lee has notched a victory. But AlphaGo has consistently won each of the games played thus far.
“I am very sorry for the powerless display”, Lee said.
Chinese Go ranking No. 1 Ke Jie, in an interview with Korean media Yonhap, sent a congratulatory message to Lee saying he hoped his played his own game. With its overall victory sealed, the AlphaGo team takes home a $1 million prize, which will be donated to STEM societies and UNICEF (among other organizations). American commentator Michael Redmond (9-dan) said: “Today’s game was another example of AlphaGo playing a very interesting, good game”.
It seems as if this will be the last board game that Google DeepMind tries to tackle in such a public forum.
“It is a huge landmark for Go, a long-anticipated moment in the West and – to a greater degree than I imagined – a shock to many people in Asia”, said Andrew Okun, president of the American Go Association.
AlphaGo uses two sets of “deep neutral networks” that allow it to crunch data in a more human-like fashion – ignoring millions of potential moves that humans would instinctively know were pointless.
South Korea’s Lee Sedol, one of the world’s best Go players, remained winless against AlphaGo, Google DeepMind’s artificial intelligence machine, after another close match in Seoul. Tweets from Demis Hassabis, the founder of Google’s DeepMind project, which is responsible for creating AlphaGo, revealed that AlphaGo made a misstep on move 79 and didn’t realize it until move 87.
The game of Go originated in China more than 2,500 years ago.
In the game, players take turns to place black or white pieces, called “stones”, on the 19-by-19 line grid, to aim to capture the opponent’s stones by surrounding them and encircling more empty space as territory. Instead, AlphaGo combines two methods of analysis, drawing upon a database of well-played Go games as well as Monte Carlo simulations to evaluate board positions. It taught itself how to go from an amateur player to world champion in less than a year. So as I play the fourth and fifth match I do ask that you continue to show interest and follow what happens. He suggested that the program had got a lot stronger and more sophisticated after its 5-0 win in October against European Go champion Fan Hui.
Advertisement
Google executives say that Go offers too many possible moves for a machine to win simply through brute-force calculations, unlike chess, in which IBM’s Deep Blue famously beat former world champion Garry Kasparov in 1997. “I’m quite speechless”, he said at the press conference that was livestreamed on YouTube.
