Fresh off its AlphaGo system defeating European champion Fan Hui five times in a row, Google has its sights set on livestreaming the AI’s next challenge, against the highest ranked player in the world, Lee Sedol.
To sweeten the pot, the winner will get a $1 million prize.
‘Go’ is an ancient Chinese board game that is simple to learn, but features millions of possible combinations for its pieces, which has proven challenging for previous AI researchers. Past Go robots have only been able to play to an amateur level, but AlphaGo was trained using 30 million moves from games played by human experts.
Along the way, AlphaGo even learned a few moves on its own.
So far, machines have been remarkably good at taking down champions in their respective fields. IBM’s Deep Blue, for example, has beat grandmasters at chess and another IBM creation, Watson, has thoroughly dismantled Jeopardy champions.
AlphaGo is significantly more complex than Watson or Deep Blue, so it should be worth checking out if you’re into that sort of thing.