Artificial intelligence of OpenAI development wins human world champion with 1: 1 battle of game Dota 2
An example of artificial intelligence (AI) winning against human beings is becoming common in games such as chess, go and poker. Established with the aim of safe development of AIOpenAIMade a fantastic achievement of winning the human champion with direct battle in PC games, which is required for proper judgment under circumstances more complicated than chess and Go.
Dota 2
https://blog.openai.com/dota-2/
OpenAI's AI is a human championDanil "Dendi" IshutinYou can check the situation of breaking the following movie after 12 minutes and 15 seconds.
The International 2017 - Main Event Day 5 - YouTube
Mr. Ellon Mask, one of the founders of OpenAI, in a fruitful attempt to defeat the human champion, said, "I broke humans with far more complex eSports than chess and go".
OpenAI first ever to defeat world's best players in competitive eSports. Vastly more complex than traditional board games like chess & amp; Go.
- Elon Musk (@ elonmusk)August 12, 2017
The following movie explains how OpenAI created AI to defeat the human champion.
OpenAI + Dota 2 - YouTube
OpenAI developer Greg Blockman said, "The purpose of OpenAI is to develop a secure AI," while rapidly evolving AI develops AI technology safely as AI threat theory is raised We are doing various efforts to do. Development of AI to win human beings in battle game is also done as part of that.
OpenAI has been developing AI to defeat human champions with the theme of online game "Dota 2".Multiplayer Online Battle Arena (MOBA)The game called Dota 2 is a complicated game that is required to properly perform many tasks according to the situation and it is a perfect game for developing new technologies to further develop AI technology.
The competition genre of this time is a one-on-one match called "1v1". In 1v1 it is required to predict the attack and defense of the opponent, find out the attack method behind the opponent, and execute it. 1 v 1 is not a game in which reflexes competing for the number of attacks are questioned, but rather a game with a personality brain fighting game, rather than winning a human champion in brain fighting is the purpose of AI development this time.
AI developed to win human beings is a completely self-learning type and has improved technology. While AI himself plays the game without using imitative learning or tree search, gradually increase the skill as if studying his / her style of play and going up the stairs step by step It seems that the process to go is adopted.
OpenAI's AI got through two weeks of self-learning, and it was decided to show off the game capability because the Dota 2 World Congress that gathered 20,000 Dota fans in the hall "The International 2017"
Totaling 24 million US dollars (about 2.6 billion yen) prior to the game tournament The International 2017 the highest ever amount of prize money has been betting that, OpenAI is allowed to play against a few of the top professionals of the AI was developed, fight with a human champion He said he was getting a response.
"Many professional players hoped to continue fighting the AI bot in training," Blockman revealed.
Artour "Arteezy" Mr. Babaev is one of them. There are many things to learn when reviewing the battle against AI with the movie, and it says that the battle against AI is extremely useful for skill improvement.
OpenAI, which controlled a one-on-one battle with humans, has set the next goal of adding AI to the human team with a 5 to 5 team battle.
Musk the rapid evolution of AI is to point out the danger of giving to future humanity is, again, "" but no one likes a regulation ", those on the danger to the public, such as automobiles, airplanes, food, and drugs It is regulated .AI also tweet and to what is "is regulated.
Nobody likes being regulated, but everything (cars, planes, food, drugs, etc) that's a danger to the public is regulated. AI should be too.
- Elon Musk (@ elonmusk)August 12, 2017
Related Posts: