Lee Sedol Finishes Incredible Series Vs Google DeepMind's AlphaGo
AlphaGo won the series 4-1, but Sedol went out fighting.
Lee Sedol’s marathon battle against Google’s DeepMind AlphaGo artificial intelligence is over. AlphaGo won the final match Tuesday morning after a grinding, nail-biting comeback against the master human go player.
When AlphaGo swept the five-match series 3-0, many commentators though Sedol was doomed to lose all five. But on Sunday night, Sedol did something everyone was starting to think was impossible. He beat AlphaGo, Google DeepMind’s incredible A.I. that had seemingly mastered the ancient Chinese game of Go. Sedol, a 9-dan professional, had lost the first three matches, but in the fourth, Sedol scored one for humankind, playing a brilliant move midway through the game and flummoxing the computer.
Sedol’s move was mathematically incredible, and AlphaGo was completely unprepared.
Early in game five, it DeepMind founder Demis Hassabis thought Sedol had thrown a glitch into AlphaGo’s system again.
But AlphaGo slowly inched back into the game and eventually took a slight advantage. The English commentary team — Chris Garlock, a vice president of the American Go Association and 9-dan pro Michael Redmond — thought the match was so even it would go to a count, the first of five matches to do so. But Se-dol resigned just before the end, realizing he had lost narrowly on points, no matter what.
The problem (or genius) with AlphaGo is that it keeps getting better. Sedol’s move in game four was like nothing the software had seen before, so it struggled to adapt to such a mathematically improbable move.
And even after faltering in the beginning of game five, AlphaGo still clawed out a victory against one of the world’s greatest players.
Stay tuned to DeepMind’s YouTube channel for updates and a breakdown of the match Tuesday afternoon. You can watch a 15-minute recap of game four, in which man triumphed over machine, below.
AlphaGo and Lee Sedol’s battle may have completely changed the future of both go and artificial intelligence](https://www.inverse.com/article/12620-elon-musk-says-google-deepmind-s-go-victory-is-a-10-year-jump-for-a-i). As Garlock put it after the fifth match, “I think we’ll be studying these [games] for years to come.”
