“But robots can’t apply logic and strategy!” they said.
Well, they were wrong.
Yesterday, a computer program defeated humans at Go, an enormously complicated strategy board game that’s said to be the oldest in the world. The breakthrough is especially notable because Go involves a great degree of strategy and can’t be solved by number-crunching alone. Human experts believed it wouldn’t be possible for A.I. to achieve this for another decade, which, if they’d just consulted a robot, they’d have known this forecast to be erroneous.
The algorithm was developed by Google’s U.K.-based DeepMind artificial intelligence lab, which swept European Go champion Fan Hui five games to none.
“Go is probably the most complex game ever devised by humans,” said Demis Hassabis, a Google DeepMind founder.
“It has 10 to the power of 170 possible board positions, which is greater than the number of atoms in the universe. It takes a lifetime of study to master, and that’s what makes it so fascinating for humans to play, and also such a great challenge for A.I. research.”
As the Toronto Star reports, Go has too many possible outcomes to use so-called “brute force” calculation. Instead, Google DeepMind researchers programmed the computer to combine a number of approaches from “deep learning,” which uses human-like learning methods such as pattern recognition, experience application, and image recognition to solve problems.
“We’re excited to see what we can use this technology to tackle next,” said Demis Hassabis, a Google DeepMind founder.
Perhaps England’s best chance at winning another World Cup lies in fielding a team of robots.