In the lead up to the historic match between the world’s top Go player Lee Se-dol and Google DeepMind’s artificial intelligence computer AlphaGo, there are mixed views among experts about who will emerge victorious.
Experts in the Go field -- including Lee Se-dol himself – said that the human will be the winner of the five-game match, which kicks off on March 9. The professional Go player predicted a 5-0 or 4-1 victory in his favor during a press briefing on Feb. 22.
Showing confidence in winning, Lee said, “Because the company has had only five months to improve the system since its game against Fan Hui, I don’t think it will be a very close match.” The French two dan player Fan Hui lost to AlphaGo 5-0 in October.
Other professional Go players shared similar views. Lee Chang-ho, one of the strongest modern Go players in the world, also predicted, “It will be an interesting match, but I think Lee Se-dol will win this time.”
However, experts in the AI field, including Google’s DeepMind, rated AlphaGo’s competence a little higher.
“AlphaGo must have thrown down the gauntlet to the world’s top player after completing its preparations to win,” said Kim Jin-ho, a professor, who lectures on big data at the Seoul School of Integrated Sciences & Technologies.
“The computer takes only a few weeks to learn from one million matches while a human needs more than 1,000 years,” he said, predicting “AlphaGo will win a complete victory in the five-game match.”
DeepMind’s founder Demis Hassabis also predicted that AlphaGo has a fifty-fifty chance of winning, saying, “AlphaGo’s goal is to beat the best human players not just mimic them.”
Although there are different views toward the outcome of the historic game that will take place between man and machine, one thing seems clear – the machine will ultimately be the winner.
Most professional Go players, including Lee himself, have predicted that Lee will win this time, but they remain uncertain about future wins.
Lee had said, “I may win this time but I am not sure what will happen in one to two years.”
To AlphaGo, winning or losing in the upcoming match is not important, said DeepMind’s Hassabis.
What DeepMind wants from the match is to gain data from the world’s top player and to eventually become the top in Go.
“The development speed of AI, which learns without rest or sleep, has been faster than the expectations of scholars. After all, time is on AlphaGo’s side,” said Lee Sung-hwan, a professor from the department of brain and cognitive engineering at Korea University.
“Even if AlphaGo loses at the match, it is almost 100 percent sure that the computer will win by the end of this year because it can learn even without Go records,” said Kim Jin-hyung, a professor of KAIST and head of the Software Policy & Research Institute.
Despite such claims, there are some experts who still wish to bet on humans.
“It is true that computers have already outdone humans’ memories and calculations. But, they still don’t have something humans are good at,” said Lee Kwang-hyung, a dean of KAIST Moon Soul Graduate School of Future Strategy. “Creativity -- it is still men’s domain.”
By Shin Ji-hye (
shinjh@heraldcorp.com)
