At last the computer won the professional shogi player with go and the knockout was Google's artificial intelligence technology
Computer software that breaks chess, shogi and professional players has come on but the day of defeating professional players came in the world of "go" that remained as the last fortress. Artificial intelligence software that broke down professional players "AlphaGoI developed theGoogle ResearchSo, deep learning is a technical key.
Research Blog: AlphaGo: Mastering the ancient game of Go with Machine Learning
http://googleresearch.blogspot.jp/2016/01/alphago-mastering-ancient-game-of-go.html
You can see how AlphaGo breaks down professional players in the following movie.
Google DeepMind: Ground-breaking AlphaGo masters the game of Go - YouTube
The evolving computer has demonstrated performance improvement by breaking the human in the game world. Starting with the 1992 board game ... ...
Othello in 1994 ... ....
And in 1997Garrie KasparovIBM'sdeep blueHas won, chess is also capturing.
After that, the world of shogi was also captured by computer, and the world of Go was seen as the last fort.
In the case of chess, on average 24 hands can be chosen as the next hand.
On the other hand, in the case of Go, there are nearly 200 candidate hands, and because of this complexity, it is impossible to process a huge amount of calculations even with the latest computers, which makes it difficult for Go That is why it was considered the last fortress.
It was Google that challenged this impregnable fort. He shot three times for a European champion from ChinaJu HuiI shake hands with Mr. Mr. before Mr. Fuji, Google's Artificial Intelligence LaboratoryDeep MindLed byDemis HasabisMr.
The game started.
It is the reason for the opposite situation to measure the strength of AlphaGo against the European champion and try the evolution of artificial intelligence technology.
Mr. Hui looking back to the game. "1st station, I lost"
Mr. Hasabis who saw Mr. Hui who can not hide embarrassment by being torn, looks back, "He seemed not to be ready enough."
"Because I made a mistake, Mr. Hui who thinks that if he can not win without a mistake, he will win," he said.
But the strength of AlphaGo is real.
The game was played 5 times and the result was AlphaGo's 5 wins.
Finally, AlphaGo got the title "world's first computer software in the world of go" to win a professional shogi player. However, European champion alone seems insatiable.
What AlphaGo aimed next was the treasure of KoreaLee Se ChengKudan It is a top pro who is seen as one of the world's strongest players called "The Devil of the World".
Google Research is planning a game against Iku in March 2016. If AlphaGo can defeat the demon kings of the game world, it seems that it will be evaluated as "Computer grasped the world of Go" without complaint.
About the research results of the artificial intelligence software AlphaGo, the paper is published in science journal Nature as follows.
Neural network training pipeline and architecture.: Nature: Nature Publishing Group
http://www.nature.com/nature/journal/v529/n7587/fig_tab/nature16961_F1.html
In AlphaGo, as a traditional handwriting search algorithmMonte Carlo tree searchIn addition to thatDeep learningHe said he took in technology. A hybrid strategy is adopted that selects the best hand using two kinds of evaluations, the policy network and the value network that reduces the depth of the search tree by narrowing down only the hands that have a high possibility of winning .
The present rating released by Google Research is like this.
Incorporating machine learning technology to learn computers such as deep learning as well as a conventional method relying on computing performance "to verify possible hands at high speed", until now, "To win a professional shogi player of go It will be said that Google changed the evaluation which had been said "It will take 10 years."
·bonus
In addition, Facebook, which is also focused on research on artificial intelligence, also seems to be utilizing the battle in Go to refine technology on Facebook. Through technological development of deep learning, computers seem to evolve acceleratingly.
(3) Mark Zuckerberg
https://www.facebook.com/zuck/videos/vb.4/10102619979032811/
The ancient Chinese game of Go is one of the last games where the best human players can still beat the best artificial intelligence players. Last year, the Facebook AI Research team started creating an AI that can learn to play Go.Scientists have been trying to teach computers to win at Go for 20 years. We're getting close, and in the past six months we've built an AI that can make moves in as fast as 0.1 seconds and still be as good as previous systems that took years to build. Our AI combines a search-based approach that models every possible move as the game progresses along with a pattern matching system built by our computer vision team.The researcher who works on this, Yuandong Tian, sits about 20 feet from my desk. I love having our AI team right near me so I can learn from what they're working on.You can learn more about this research here: http://arxiv.org/abs/1511.06410
Posted byMark ZuckerbergOn 26th January 2016
· Continued
At last Google's artificial intelligence "AlphaGo" and the world's strongest players "Lee Se-dul" confrontato and live on YouTube - GIGAZINE
Related Posts:
