The DeepMind AI Is Now Invincible To The Go Game According To A Great Champion

The DeepMind AI Is Now Invincible To The Go Game According To A Great Champion

Science

South Korean Lee Sedol, one of the great champions of the game of go decided to retire. The fact is that the interested party took advantage of this moment to affirm that DeepMind’s artificial intelligence (AI) was simply invincible.

AlphaGo has won everything
The UK company DeepMind, owned by Alphabet (Google) since 2014, has developed AlphaGo, a high-performance AI. It is regularly talked about in the media. In 2018, one study claimed that this AI could support ophthalmology. Indeed, AlphaGo has been able to read the image of an eye and identify with a high success rate the presence of no less than fifty eye diseases.

However, the AI ​​has mainly distinguished itself in the practice of the game of go. However, you should know that this board game is considered one of the most difficult to understand for an artificial intelligence. And yet, AlphaGo managed to beat defending European champion Fan Hui in 2015, one of the best players in the world, Lee Sedol, in 2016 and finally, world champion Ke Jie in 2017!

The retirement of a champion
In a statement issued on November 27, 2019, South Korean Lee Sedol announced that she has retired. He stated that the progression of the AI ​​in the game of go made him aware that he no longer had the weapons to compete. Lee Sedol believes that even if he continues to progress and becomes the best player in the world, the AI ​​will continue to beat him because it is much too strong.

It must be said that the man has trouble digesting the game of 2016, lost on the score of 4 to 1. Lee Sedol had fallen from very high when he was considered one of the best players in the world. And yet, it remains to this day the only human to have removed a round at the IA in a game of go!

For DeepMind, these reigning victories sounded the consecration. Indeed, the game of go has long been considered inaccessible due to its complexity. Now the AI ​​is at the top of the ladder. Above all, an improved version of AlphaGo appeared in 2017: AlphaGo Zero. However, it has managed to beat the old version 100 times in a row! In other words, the decision of Lee Sedol is perfectly understandable because, obviously, no human will be able to compete again.