Monday 07 March 2016 5:44 pm

DeepMind’s AlphaGo versus Go world champion Lee Sedol: Everything you need to know about the man versus machine showdown

Artificial intelligence is about to face its biggest test to date.

The AI software AlphaGo created by British-based Google-owned DeepMind will face its most challenging human foe yet – the champion player of the chess-like ancient Chinese game Go, Lee Sedol.

Man and machine will meet for a five set game in Seoul starting this week.

Representing the human corner is Sedol, the Roger Federer of Go according to one DeepMind engineer. The 32 year-old South Korean and champion of the highest rank with 18 international titles under his belt has expressed confidence. "I'll win the game by a landslide – at least this time," he boasted to reporters last month ahead of the match.

Representing the march of AI is AlphaGo, the first self-taught machine to master the game to such an extent that it can beat more than the amateur player. In a landmark publication earlier this year, DeepMind revealed that it had already beaten European champion Fan Hui – a feat founder Demis Hassabis said was previously believed to be a decade away.

AI has already mastered several games – famously chess, when IBM's Deep Blue beat champ Garry Kasparov in the 90s, but also more recently DeepMind mastered Atari games – but Go is the Holy Grail.

The 2,500 year-old game Go presents one of the "grand challenges" of AI due to its complexity. Within that complexity, there are still rules, however.

"AI researchers were always very interested in games, because they were extremely complex, huge number of possible positions within games were available, yet their simple in a certain way," said IBM research scientist Murray Campbell. "They're simple in that the moves are well defined, the goals are well defined, and so you don't have to solve everything all at once."

At stake is not just the $1m prize money (what does AI, or Google for that matter, need it for), but scientific and technological progress.

“Go is the most profound game that mankind has ever devised," said Hassabis ahead of the tournament. "The elegantly simple rules lead to beautiful complexity. Go is a game primarily about intuition and feel rather than brute calculation which is what makes it so hard for computers to play well."

If AlphaGo wins against Sedol in what's being called its truest test yet, it would represent the most human-like machine yet created.

“The Lee Sedol AlphaGo match is not just about AI challenging human in a more complex game, Go, after beating human at chess," said Jeong Jaeseung, professor of of bio and brain engineering at Korea Advanced Institute of Science and Technology.

"Unlike the past when AI calculated the probabilities and patterns for all moves, AI is challenging human with humanlike reinforcement learning and strategy derived from deduction. The match, regardless of the result, will become a landmark in the history of AI.”