Alpha Go

The AlphaGo research project was formed around 2014 at DeepMind, the London based company acquired by Google, to test how well a neural network using deep learning could compete at Go, an ancient board game long considered one of the hardest challenges for artificial intelligence. Go is played on a 19 by 19 board, and its number of possible positions is far larger than in chess, the game IBM's Deep Blue had mastered in 1997. AlphaGo uses Monte Carlo tree search guided by neural networks trained on both human and computer play.

In October 2015, AlphaGo defeated European Go champion Fan Hui five…