Watch OpenAI Bots Destroy Top Human Players in Dota 2

Adjust Comment Print

These are all people who now or previously play Dota 2 professionally - they're in the 99.95th percentile of human players.

It only lost a game series when it had its "heroes" selected by audience members watching the action unfold, and those selected characters severely handicapped the OpenAI Five.

Prior to the Benchmark Game a team of volunteers for the audience played a public match against OpenAI Five.

OpenAI bots played a match (with some significant alterations) against current pro player David "MoonMeander" Tan, former pros Ben "Merlini" Wu, William "Blitz" Lee and Ioannis "Fogged" Lucas and caster Austin "Capitalist" Walsh.


A few years ago, many researchers thought it was impossible for AI systems to get so good at complex team-based games like Dota 2, but here we are.

The bots came from OpenAI, a non-profit AI research company co-founded by Elon Musk.

OpenAI's initiatives with Dota 2 began in March previous year. Playing against bots and one mistake could cause end of the match.

Through this reinforcement learning, the networks learn to cooperate with each other, build "team spirit" and take down in-game opponents.


The latest victory for OpenAI Five is a positive step towards the firm's goal of competing against the best proteamat The International which is the largest annual Dota tournament in the world. In one week of implementation, we crafted a fake frame for each of the 11 million possible team matchups and wrote a tree search to find OpenAI Five's optimal draft.

After losing the first three games of Go to Google's AlphaGo, Lee Sedol, a professional player ranked second in worldwide titles, deployed a highly unusual strategy to win the fourth game, inspired in part by the idiosyncratic algorithmic playing style of the AI that he had observed during the previous three games. Well, the AI platform developed by OpenAI would like to have a word, humans.

"It's hard to know what kind of progress you're making if you're just making progress in simulators", Jack Clark, who works at OpenAI, said in an interview with Axios.


Comments