Yahoo Autos Tonight: Watch an AI Take On One of the World's Best Go Players
In late January, Google's AlphaGo deep neural network was able to best European Go champion Fan Hui in a full 19 by 19 game grid. It was an amazing leap forward for neural networks. While it's been generally accepted that even the world's best chess players can be beat by a chess program running on commercially available hardware since the mid-2000s, Go has resisted the brute force methods that allowed computers to supplant humans as the best chess players on Earth.
This is mainly due to the complexity of Go. On average, there are about 20 things a player can do in chess, versus the 200 things a player can do in Go. Play this scenario out a few times, and the number of potential plays on a 19 by 19 board quickly rises into numbers you need to have to notate like so: ~2.082 × 10^170 (i.e., a 2 followed by 170 zeroes).
