Lee Se-dol, Not As Confident After Understanding the AlphaGo's Algorithm

Hello, I'm back with the latest news on Lee Se-dol vs. AlphaGo.

For the first time visitors, please find the below related articles for more information on the Lee Se-dol vs. AlphaGo.

Related Articles

▶ [Alphago vs. Lee Se-dol] Who is this Lee Se-dol guy?

▶ A Big Heads Up in Three Weeks : Korean Baduk(Go) Grandmaster vs. AlphaGo

It is about time guys. The big match is upcoming in several hours. The exact time would be March 9 at 1:00 pm in South Korea, or March 8 at 8:00 PM Eastern Standard Time. So, don't miss out on this. However, it is a pity that many of us are unfamiliar with the game of Baduk or Go. Oh well, what can we do. Let's just delve into what Lee Se-dol said yesterday at the before-the-match press conference, where 300 reporters gathered from all over the world.

This guy doesn't look any nervous!

Previously, Lee Se-dol was very confident and said several times to the media that he will beat the software by five to nothing. But, in yesterday's interview, he was more cautious and conservative about the prediction of the result. The reason for the change in his attitude going into the game was that he now understands how the algorithm of AlphaGo works and it simply exceeds what he imagined. He now thinks he can lose a game or even two. 

Se-dol believes that what differentiates human players from software is the intuition. Intuition is a very vague term, but here it means the ability to branch out unnecessary possibilities of outcomes and focus on the important branches. The previous programs viciously searched millions and trillions of outcomes just to find one good move, and it is just not possible to brute force the game of baduk because of the reason I explained before, it has close to infinite possibilities of outcomes. 

But after the technical session with DeepMind's CEO Demis Hassabis, Se-dol's belief on the google's program changed. He now thinks that AlphaGo just might be able to imitate human's intuition to some extent. Hassabis also didn't forget to mention some of his program's strengths, which are its endurance and inability to fear. Despite the reduced confidence and toned-down atmosphere, Se-dol ended the interview with a positive note. 

"AlphaGo could have imitated the human's intuition and feel. But my feeling is that those qualities could not have been implemented to 100%. That is why a human still has an advantage over a machine."

I'll be back with more of Se-dol vs. AlphaGo as the game goes on. Thanks.