OpenAI bots now able to defeat the humans in DOTA 2 team matches: Breakthrough for artificial intelligence



[ad_1]



OpenAI has developed robots powered by artificial intelligence that are now able to beat humans in "Dota 2" matches of 5 against 5. Known as the OpenAI Five, the bots will face the best players of "Dota 2" in the world this year.
( OpenAI )

San Francisco-based OpenAI has developed bots that are now able to beat humans in 5-on-5 team games on a popular strategy game Dota 2, marking a milestone for research on artificial intelligence.

Al algorithms usually work independently, so this is considered a major breakthrough because OpenAI has been able to run algorithms with each other by beating their human opponents.

OpenAI Bots Beat Humans In 'Dota 2' Team matches

OpenAI, an artificial intelligence research start-up supported by Elon Musk, CEO of Tesla and SpaceX, has launched an AI-powered robot last year at the International, the biggest Dota 2 tournament. The bot faced and won against Danil Ishutin, better known as Dendi and considered one of the best Dota 2 players in the world, in 1-on-1 matches.

Near a year later, OpenAI comes back with the algorithmic team A that the research lab calls OpenAI Five. Similar to last year's bot at The International, each OpenAI Five algorithm uses a neural network to learn how to play Dota 2but also by teaming up with his AI team members.

In an official blog post, OpenAI revealed that the OpenAI Five play 180-year games against themselves each day. This intense training allowed the bots team to start defeating teams of amateur human players Dota 2.

OpenAI will close its loop in a few months, when the OpenAI Five will challenge the best Dota 2 Worldwide teams in the International this year.

Artificial intelligence in video games

Why do researchers teach artificial intelligence to play video games, among all possible tasks? The reason is simple: if AI systems are able to learn the skills necessary to play video games, they will be able to use these skills to solve real problems. Some of these problems, such as managing groups of people or determining the quickest solution, are similar to what players encounter in video games.

It also helps to control Dota 2 is a difficult task for AI. Chess and Go, board games that have seen their fair share of AI apps, can end in less than 40 and 150 shots, respectively. Meanwhile, the robots must perform 20,000 movements in 45 minutes Dota 2 meet.

"As long as the AI ​​can explore, it will learn, with enough time," Greg Quartz, Technology Manager at OpenAI, told Quartz. If artificial intelligence can learn to play Dota 2 good enough to beat humans, so the potential for technology is unlimited.

The key step in the collaboration between AI systems, as evidenced by the OpenAI Five, opens a wide range of applications for this technology. AI algorithms may be able to team up to accomplish tasks even faster, and may even work with humans to achieve goals.

See now:

30 Gadgets and Gifts Tech for Father's Day 2018 Daddy Will Think to be Rad



© 2018 Tech Times, All Rights Reserved. Do not reproduce without permission.

[ad_2]
Source link