Go champ recalls defeat at hands of 'calm' computer

Go is something of a Holy Grail for AI developers, as the ancient Chinese board game is arguably more complex than chess
Go is something of a Holy Grail for AI developers, as the ancient Chinese board game is arguably more complex than chess

Last October, Fan Hui was beaten by a computer at the ancient board game of Go that is not only his passion but also his life's work.

This week, the 35-year-old European champ will referee as his vanquisher, Google's AlphaGo programme, faces the world's human Number One in a battle for Go supremacy.

"I was the first professional Go player to be beaten by a computer programme. It was hard," Fan told AFP ahead of AlphaGo's five-day duel against Go world champion Lee Se-dol.

"At the beginning of the match, I never imagined I could lose."

Said to be the most complex game ever designed, with an incomputable number of move options, Go requires something akin to human "intuition" to prevail.

Game-playing is a crucial measure of AI progress—it shows that a machine can execute a certain "intellectual" task better than the humans who created it

Fan is being held to secrecy about AlphaGo's playing style and his expectations for the outcome of the match.

But he did insinuate that Lee will have his work cut out for him if he wants to take home the $1-million (908,000-euro) prize money.

"He will face a machine that is much stronger than the one that played against me," said the Chinese-born, Bordeaux-based Go teacher.

The marathon match, to be played over five days, is seen as a test of how far Artificial Intelligence (AI) has advanced.

What makes AlphaGo special is that it is partly self-taught—playing millions of games against itself after initial programming to hone its tactics through trial and error.

Lee Se-Dol (R), a legendary South Korean player of Go poses with Google Deepmind head Demis Hassabis during a press conference a
Lee Se-Dol (R), a legendary South Korean player of Go poses with Google Deepmind head Demis Hassabis during a press conference ahead of the Google DeepMind Challenge Match in Seoul on March 8, 2016

Feeling better

"Five months later, I feel a lot better," said Fan of his defeat, recounting how he became stressed during his own five-day face-off with AlphaGo in London, lost confidence, and started making errors.

"The computer, on the other hand, played calmly," said Fan. "It didn't make any big error. It did not stress out at all."

Fan is excited about witnessing his human and computer rivals face off, describing the contest as "perhaps the most important event for Go since the game's invention" some 3,000 years ago.

Until Fan's defeat, computer mastery of the game was thought to be at least a decade off.

Go reputedly has more possible board configurations than there are atoms in the universe, a much higher challenge for AI than say the game of chess, in which computers already outplay humans.

Initially confident of winning by 4-1 at worst, Lee's courage seemed to have started waning by Tuesday.

He told reporters in Seoul the system seemed to work "far more efficiently" than he thought at first, and "I may not beat AlphaGo by such a large margin".

In the end, added Lee, it seems "inevitable" that AI will defeat humans at Go.

"But robots will never understand the beauty of the game the same way that we humans do."

Lee is one of the greatest players in modern history, having won 18 international titles.

He is ranked at the top of the nine-level scale for professional players, far above Fan who is at the second level.

© 2016 AFP

Citation: Go champ recalls defeat at hands of 'calm' computer (2016, March 8) retrieved 26 April 2024 from https://phys.org/news/2016-03-champ-recalls-defeat-calm.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.

Explore further

Game over? New AI challenge to human smarts (Update)

8 shares

Feedback to editors