Humanity Notches First Go Victory Against Google's AI Champion Go player Lee Sedol picks up his first win against Google's not-so-unbeatable AlphaGo after three losses.
By David Murphy
This story originally appeared on PCMag
Like the Terminator, Google's AI creations can be defeated.They might get the better of you sometimes, if not most of the time, but they aren't invincible. At least not when it comes to the maddeningly complex game of Go.
Though Google's AlphaGo program already notched three wins against champion Go player Lee Sedol, which gives it the win for the five-game series between Lee and AlphaGo, the human player squeezed a victory out of a strong counterattack from Google's system early in the game. Lee nailed a big move on White 78, sending AlphaGo into what could be best described as a downward spiral of mistakes -- the first being on its immediately next move, 79, which AlphaGo finally realized was a problem around move 87.
"Lee Sedol is playing brilliantly! #AlphaGo thought it was doing well, but got confused on move 87. We are in trouble now," tweeted Demis Hassabis, founder of Google's artificial intelligence division, DeepMind.
"When I say 'thought' and 'realisation' I just mean the output of #AlphaGo value net. It was around 70% at move 79 and then dived on move 87," he added.
The exciting aspect of Lee's victory today is twofold. Not only has he notched up his first win against the AI, but it also suggests that Sedol is starting to figure out strategies that might work against AlphaGo in the next (and final) match in the tournament on Tuesday.
"It seems Lee can now read AlphaGo better and has a better understanding of how AlphaGo moves. For the 5th match, it will be a far closer battle than before since we know each better. Professional Go players said that they became more interested in playing Go after witnessing AlphaGo's innovative moves. People started to rethink about moves that were previously regarded as undesirable or bad moves. AlphaGo can help us think outside of the box in Go games," said Korean commentator Song Taegon, as reported in a Google blog post.
Of course, that's also assuming that AlphaGo isn't going to learn from its mistakes today and use that knowledge to thwart Lee's plans for the fifth game -- machine learning and a deep neural network is the main reason why AlphaGo is so good at Go to begin with.
For Lee, even racking up one victory against AlphaGo is an accomplishment, especially since AlphaGo is now 9–1 against all the Go champions it has faced. "I've never been congratulated so much just because I won one game," said Lee, responding to the huge applause he received in today's post-game press room, as reported by the The Verge.
