The AlphaGo, a system of neural networks developed by Google-owned DeepMind, defeated fabled Go player Lee Se-dol Wednesday in a landmark competition of man versus A.I. The first in a series of five matches, the game took place in Seoul and lasted about three and a half hours before Se-dol forfeited.

Developed in ancient China, Go is an extremely complex game that tasks players with moving stones throughout a board in an effort to control as much of the playing space as possible. The game might sound a bit like chess, but the level of instinct and calculation required to make the moves is much greater.

“I was very surprised,” Lee told the Verge after the match. “I didn’t expect to lose. [But] I didn’t think AlphaGo would play the game in such a perfect manner.”

DeepMind founder Demis Hassabis took to Twitter during the match, and displayed his joy following its conclusion.

“I don’t regret accepting this challenge,” said Lee, who is playing for $1 million. “I am in shock, I admit that, but what’s done is done. I enjoyed this game and look forward to the next. I think I failed on the opening layout so if I do a better job on the opening aspect I think I will be able to increase my probability of winning.”

AlphaGo functions using an advanced tree search algorithm and two different deep learning neural networks. The “policy network” helps AlphaGo to make its next move. The “value network” maintains a running calculation of which player will win the game. Using the networks, AlphaGo can identify 30 million moves that can be made by its human competition.

The machine’s win over Se-dol follows its five-game sweep of three-time European Go champion Fan Hui in 2015.

The final four of the games in the series will take place Friday, Saturday, Sunday, and Tuesday in Seoul, meaning that they would take place at 11 p.m. ET the previous night.

[Via The Verge]