AlphaGo-Go-Go


The game of “Go” originated in China more than 2,500 years ago. It is considered to be one of the most complex board game on earth, the number of possible games is vast, 10170 . That's 10 to the power of 170! It has been proven to be “PSPACE-hard” in 1978 by Robertson and Munro.

If you type four words “Can a computer beat” in google, google will show a lot of interesting suggestions based on the most user searches phrases in the entire world.

The reason why “Can a computer beat a human at go” is in third place, is simply because in March 2016, there was a game hosted in Seoul, South Korea between a computer program and humans. The program was invented by engineers from Google, named AlphaGo. And the human - Lee Sedol, is a professional Go player of 9-Dan, he also ranks second in international titles.

It wasn't the first time AlphaGo has played a match against professional Go players. In October 2015, AlphaGo won 5-0 in a formal match against the reigning 3-times European Champion, Fan Hui, to become the first program to ever beat a professional Go player in an even game.

Before the match, most people expected Lee to win, or at least have the upper hand. Surprisingly AlphaGo won the first 3 matches, although Lee fought back and toke the forth game, AlphaGo ends it with a 4-1 victory.

The A.I. (artificial intelligence) of AlphaGo is better than most human brain, although it is not perfect as it has lost 1 game. This leads to the question, how far can A.I. go?

Science-fiction novels or movies like “I-Robot”, “Terminator” and “The Matrix” have all tried to explore how A.I. will eventually be able to out-smart humans. It may seem more fiction than science at the moment. However, human behavioral learning is at its all time highest in terms of subject participation from search engines like Google and A.I. assistants like Siri.

Thanks in a great part by the internet and smartphones, every second, tens of thousands of queries are being sent to Google and Apple. A.I. learns that users of different countries, race, age, gender or religion have vastly different preferences. From this the algorithm behind the A.I. predicts what the user is looking for and react accordingly.

Just like the Google search we did at the beginning of this article.

Predicting user pattern was the reason behind complex A.I. developments from companies like Google or Apple, but they have advance greatly since then. Meant to assist humans in their everyday lives, A.I. now predicts behaviors and requests long before users know they have them.

Such intelligence of calculating many steps ahead is what makes a good go/chess player, but at the same time doesn’t this all sound eerily familiar?

You have been warned.

https://deepmind.com/alpha-go.html

https://en.wikipedia.org/wiki/Go_(game)

https://en.wikipedia.org/wiki/Lee_Sedol

http://www.koreatimes.co.kr/www/news/tech/2016/03/325_199973.html

(Photo Credits: Buster Benson)

Tags: