AlphaGo beats Go champion Lee Sedol

A computer program developed by the London-based artificial intelligence company DeepMind has beaten one of the world's best players of the ancient board game Go.

AlphaGo won the fifth and final game of its match against the South Korean champion Lee Sedol in Seoul on 15 March, taking the series by four games to one.

A landmark for artificial intelligence

Go, which originated in China more than 2,500 years ago, is played on a board of 19 by 19 lines, with players taking turns to place black and white stones. The aim is to surround more territory than the opponent.

Although the rules are simple, the game is extraordinarily complex. There are more possible positions on a Go board than there are atoms in the universe, making it impossible for a computer to calculate every possible move.

Computers mastered chess in 1997, when IBM's Deep Blue beat the world champion Garry Kasparov. But many experts believed it would be a decade or more before a computer could beat a top professional at Go.

How AlphaGo works

AlphaGo uses a technique known as deep learning, in which computer systems loosely modelled on the brain learn from large amounts of data. The program studied millions of moves from games played by human experts, and then improved by playing against itself.

It combines this with a method of searching through possible moves, allowing it to judge which positions are strong without calculating every possibility.

A British success

DeepMind was founded in London in 2010 by Demis Hassabis, Shane Legg and Mustafa Suleyman, and was bought by Google in 2014. Hassabis, a former child chess prodigy, has said the company's aim is to use artificial intelligence to help solve some of the world's biggest problems.

Lee's victory

Lee, who has won 18 international titles, managed to win the fourth game, finding a move that the program did not anticipate. He said afterwards that he had underestimated AlphaGo.

Millions of people in Asia watched the match live, and it has attracted huge interest around the world.

DeepMind says the techniques used in AlphaGo could one day be applied to problems in science and medicine, such as analysing medical scans or modelling the climate.