AlphaGo

AlphaGo
Developer(s)Google DeepMind
TypeComputer Go software
Websitedeepmind.com/research/highlighted-research/alphago

AlphaGo is a computer program that plays the board game Go. It was made by DeepMind Technologies (Google affiliate). This program became famous due to the victories against professional players.

Many new technologies were used to create AlphaGo, including deep learning, optimization, and the Monte Carlo algorithm.

Powered versions

After the release of AlphaGo, DeepMind Technologies has made powered versions such as the AlphaGo Zero and the AlphaZero. AlphaZero is a self-taught program. This means that it became powerful without human guidance.

Details

The following table is the summary of AlphaGo achievements (including its variants).

Strength of AlphaGo and its variants
VersionsHardwareElo ratingDateResults
AlphaGo versus Fan Hui176 GPUs,distributed3,144Oct 20155:0 against Fan Hui (professional player)
AlphaGo versus Lee Sedol48 Tensor processing units (TPUs), distributed3,739Mar 20164:1 against Lee Sedol (former Korean & world champion)
AlphaGo Master4 TPUs, single machine4,858May 201760:0 against professional players
AlphaGo Zero (40 block)4 TPUs, single machine5,185Oct 2017100:0 against AlphaGo version that defeated Lee Sedol

89:11 against AlphaGo Master

AlphaZero (20 block)4 TPUs, single machine5,018

Dec 201760:40 against AlphaGo Zero (20 block)

Rivals

After the appearance of AlphaGo, several research groups have created computer Go programs with similar technical viewpoints.

Darkforest

This was made by Facebook. The source codes are available on GitHub.

DeepZenGo

This was made in Japan. Nihon Ki-in was also involved in its research and development.