Following four losses, one of the world's top Go players - Lee Se-dol - has beaten Google DeepMind's AlphaGo program. Lee Se-dol, who has been described as the Roger Federer of Go, has so far only managed to beat this AI once out of his four played games, so overall, AlphaGo has already won the set.
Back in October, AlphaGo played against and beat the three time European Go champion Fan Hui, winning all five games. It also looked like Lee Se-dol would lose all five of his games as well but managed to turn AlphaGo's winning streak on its head by causing the AI to make a mistake that it couldn't recover from.
The AlphaGo AI program is different from "expert" systems which use hand-crafted rulesets. AlphaGo instead uses machine learning techniques to work out how to improve its playing style, essentially playing games of Go over and over again until it got to a state where it can play competitively against world class players.
Developing an AI for Go was considered to be a difficult task because of the immense amount of positions a player can make. In Chess, the amount of moves a player can make is around 20 for the average position (current state of the board), in Go however it's around 200. In Go there are many positions the board can be in. In fact, there are more positions than there are atoms in the Universe - or 2.08×10^170.
Source: BBC News | Image: Artificial Intelligence via Profit_Image / Shutterstock
8 Comments - Add comment