Deep Neural Networks and Igo

The game of Go has long been viewed as the most challenging of classic games for artificial intelligence owing to its enormous search space and the difficulty of evaluating board positions and moves. Here we introduce a new approach to computer Go that uses ‘value networks’ to evaluate board positions and ‘policy networks’ to select moves. These deep neural networks are trained by a novel combination of supervised learning from human expert games, and reinforcement learning from games of self-play. Without any lookahead search, the neural networks play Go at the level of state-of-the-art Monte Carlo tree search programs that simulate thousands of random games of self-play. We also introduce a new search algorithm that combines Monte Carlo simulation with value and policy networks. Using this search algorithm, our program AlphaGo achieved a 99.8% winning rate against other Go programs, and defeated the human European Go champion by 5 games to 0. This is the first time that a computer program has defeated a human professional player in the full-sized game of Go, a feat previously thought to be at least a decade away.

If you click the Affiliations tab on the published abstract you’ll see this was a Google joint: Google DeepMind, to be precise. 1

We’ve mentioned before that Google certainly seems to have replaced the mythic Bell Labs of yore – this news only solidifies that position. Just ask FB’s Zuckerberg, who posted this hilariously out-of-date item on Tuesday.

AlphaGo will be set against Go’s winningest player in March, Lee Sidol.

Deep Neural Networks and Igo

Show 1 footnote

  1. Though to be fair, the original Deep Mind founders Demis Hassabis and Mustafa Suleyman sold their company to Google just over a year ago.

Something to say...?