The world champion, who has already won 18 international titles in his career, was found on Tuesday in Seoul in five games by 280 trains. Four times lost the 33-year-old Lee against the self-learning, constant improvements to software. Only on Sunday in the fourth game he kept the upper hand.
AlphaGo and Lee played for a prize money of one million dollars. Google wants the sum now donate, among others, the UN Children’s Fund UNICEF.
Lee firmly believed in the victory against the machine
Before the first game had the world champion still tinted victory sure to win hands down against the software. The even most experts had expected because the game Go is much more varied than chess and therefore programming is extremely difficult. But the pitch has 361 boxes, chess only 64 fields. The game ways are so much greater. But Google’s Deep Mind seems to have succeeded, prepare the computer optimally for the match.
Á
the game man vs. machine was transferred to south Korea live on television – even in the U- and S-Bahn
These also served a match in October last year against the European champion fan Hui, the AlphaGo 5. 0 also clearly decided for themselves and from which the computer has apparently learned a lot
AlphaGo says trains his opponent ahead
AlphaGo has two neural networks with millions of compounds. Thus, the computer can “think” and assess the likely features of his opponent and the most likely train predict.
Á
Go world champion Lee Se-dol (l.) gartuliert Demis Hassabis (r.), co-founder of the software company Deep Mind, to victory. Go is the most difficult game in the world
But he himself seems then but to react differently than a man. Lee Se-dol said in Seoul that his computer had always surprised with features that would make a man.
Click here to read our report on the first few games of the match.
No comments:
Post a Comment