Lee Sedol versus AlphaGo

The tournament that marked a turning point in AI history.

Primer

Described as a landmark event in the history of AI, Lee Sedol's matches against AlphaGo in 2016 marked the first time an AI defeated an elite player in the difficult and ancient board game of Go.

AlphaGo was developed by Google DeepMind to play the game, which experts say was invented 2,000 years ago in China. The game is similar to Chess, but with far more possible moves and variations of play.

Lee, described as a national hero in South Korea, was incredibly confident ahead of the tournament, saying he expected to win all five scheduled matches. Watching previous games in which AlphaGo beat other professionals, many thought it wouldn't be able to beat Lee.

"Lee Sedol will have an easy win if AlphaGo's level is the same as what was shown against Fan Hui," Yoo Chang-hyuk, an elite player, said in reference to AlphaGo beating the European champion.

Ahead of the matches, AlphaGo's creators were less certain.

"Although we have programmed this machine to play, we have no idea what moves it will come up with," DeepMind research scientist Thore Graepel told Wired. "Its moves are an emergent phenomenon from the training.

AlphaGo ultimately beat Lee 4-1.

In the first game, Lee said he made a fatal error responding to AlphaGo's style of play -- Lee tested AlphaGo with an odd move. The agent's respons was "Accurate and efficient," according to a commentator, who described Lee's test as a mistake.

In the second game, Lee described AlphaGo's play as "perfect."

By the third game, AlphaGo's play was becoming genuinely convincing.

"AlphaGo won so convincingly as to remove all doubt about its strength from the minds of experienced players. In fact, it played so well that it was almost scary," David Ormerod wrote at the time.

Lee, despite all odds, won the fourth.

But his victory was short lived. AlphaGo won the fifth, though many describe it as a close game. Demis Hassabis, AlphaGo's creator, said that it made a mistake early on.

In statements after the second game, Lee described the difficulty and exasperation he felt playing against the bot.

"Normally, you can sense your opponent's breathing, their energy. And lots of times you make decisions which are dependent on the physical reactions of the person you're playing against," he told the BBC.

"With a machine, you can't do that."

In the aftermath of the games, the Atlantic quoted Andy Salerno, who posted the following message on Reddit:

AlphaGo isn’t a mysterious beast from some distant unknown planet. AlphaGo is us. AlphaGo is our incessant curiosity. AlphaGo is our drive to push ourselves beyond what we thought possible. Lee should feel no shame in his losses, for AlphaGo could never demonstrate its abilities—our abilities—if Lee were not there to challenge it.