Man falls to machine in final Go game
South Korean fails to end five-game tournament with another win, after beating supercomputer on Sunday
By Chang May Choon, South Korea Correspondent In Seoul, The Straits Times, 16 Mar 2016
When it came down to the fifth and final round of man versus machine, South Korean Go master Lee Se Dol was still no match for the Google supercomputer that had already beaten him three times.
Hopes were high for Mr Lee Se Dol, 33, one of the world's most brilliant minds in the ancient board game; he had beaten AlphaGo just on Sunday and proclaimed that he had discovered two of the computer programme's weaknesses.
But the 18-time world champion surrendered after 280 moves yesterday, unable to gain any more ground against Google's artificial intelligence program at the end of a closely fought, five-hour session.
At a press conference later, a dejected Mr Lee - a stark contrast to his beaming self two days earlier - said he felt sad he was unable to end the tournament with another win.
"I thought I gained an advantage in the beginning of the game, but eventually I lost because I was not good enough," he added. He also thanked those who cheered for him, adding that this experience will make him a better player.
Mr Demis Hassabis, founder of the Google company DeepMind that created AlphaGo, called the last match "the most mind-blowing game so far". He added that AlphaGo made a mistake early in the game but was able to recover and narrow the margin.
South Korean fails to end five-game tournament with another win, after beating supercomputer on Sunday
By Chang May Choon, South Korea Correspondent In Seoul, The Straits Times, 16 Mar 2016
When it came down to the fifth and final round of man versus machine, South Korean Go master Lee Se Dol was still no match for the Google supercomputer that had already beaten him three times.
Hopes were high for Mr Lee Se Dol, 33, one of the world's most brilliant minds in the ancient board game; he had beaten AlphaGo just on Sunday and proclaimed that he had discovered two of the computer programme's weaknesses.
AlphaGo victory: what does it mean for the future of AI? – vid...
Google’s supercomputer AlphaGo defeated world Go champion Lee Sedol 4-1.Have humans taken the 2,500-year-old game as far as they can? Is it time for computer intelligence to take over?
Posted by The Guardian on Tuesday, March 15, 2016
But the 18-time world champion surrendered after 280 moves yesterday, unable to gain any more ground against Google's artificial intelligence program at the end of a closely fought, five-hour session.
At a press conference later, a dejected Mr Lee - a stark contrast to his beaming self two days earlier - said he felt sad he was unable to end the tournament with another win.
"I thought I gained an advantage in the beginning of the game, but eventually I lost because I was not good enough," he added. He also thanked those who cheered for him, adding that this experience will make him a better player.
Mr Demis Hassabis, founder of the Google company DeepMind that created AlphaGo, called the last match "the most mind-blowing game so far". He added that AlphaGo made a mistake early in the game but was able to recover and narrow the margin.
The five-round match is viewed as a watershed moment in the ongoing man versus AI debate. Mr Lee's defeat came a decade earlier than predicted, surprising many Go professionals who thought computers generally lack the human intuition and creativity that are key to winning.
But the odds were stacked against the world's most formidable player. AlphaGo is a self-learning software backed by over 1,200 high-powered processors that allow it to make 100,000 computations per second - a thousand times more than a human brain. It is also able to mimic human thought processes and has stored and analysed over 80 million case studies.
But the odds were stacked against the world's most formidable player. AlphaGo is a self-learning software backed by over 1,200 high-powered processors that allow it to make 100,000 computations per second - a thousand times more than a human brain. It is also able to mimic human thought processes and has stored and analysed over 80 million case studies.
Mr Lee's victory on Sunday was seen as a glimmer of hope in a future world expected to be dominated by machines. He made his breakthrough with a strategic move hailed as "God's touch", which AlphaGo did not realise until it was too late. JoongAng Ilbo newspaper said his ability to spot AlphaGo's weaknesses is "testament to the fact that by far the most powerful machine still has a long way to go before it can truly match the ingenuity of the human mind".
University student Kim Yun Ah, 25, was among those rooting for Mr Lee. "I'm really amazed it's possible to win a match against artificial intelligence. I heard that AlphaGo is the equivalent of the brains of around 1,200 people, so that means Lee Se Dol fought against 1,200 people simultaneously," she said.
Mr Lee took home US$150,000 (S$207,000) for participating and an additional US$20,000 for winning a game. DeepMind, the company behind AlphaGo, will donate the prize money of US$1 million to charities including the United Nations' children's welfare agency, Unicef.
No comments:
Post a Comment