Google robot beats GO champion


Artificial Intelligence (A.I.) has made progress in the past years. The machine intelligence division at Google, DeepMind, has been focused on developing learning algorithms, which can be applied into a computer to compete with humans in games such as chess or jeopardy.

The 2,500 year old strategic Chinese game of Go is known for the endless alternatives a player can take when playing the game. The english translation for the game is ‘encircling game’.

Go is regarded as less predictable than chess. This is a reason why Go fans and grandmasters regard it as impossible for a machine to be able to calculate the best maneuvers to beat any skilled player.

Mathematical applications are evident in Go, but players consistently emphasize that adaptability and natural insight are also highly important, traits that A.I. still needs research and development to perfect.

Researcher at Google, led by A.I. specialist Demis Hassabis, took on the challenge and developed a strategy calculating machine called AlphaGo.

Lee Sedol from South Korea is regarded as a Go expert due to his masterful strategy and countless awards in Go competitions across the Asia-Pacific region. Out of five games, AlphaGo won four games while Lee Sedol won just one game. It was clear to the audience of Go fans that a machine has actually defeated a GO champion.

A.I.’s growing improvements have presented themselves on the national stage. Computers began competing and beating players at chess as early as 1997.

On the Union campus, the Go Club has taken note of these events. President of the GO club, Matt Goff ’17 remarked, “It is a melancholy that the game has lost some of its character as the ultimate strategic challenge.

At the same time it is exciting that humans have created something so advanced. It’s reassuring that the news has brought many new people to the game.”

Google’s DeepMind plans to use the experiential testing as collected data for further research and improvements on AlphaGo.

The progress that A.I. has achieved thus far will be expanded by Hassabis and his research team.


Leave a Reply