The company’s Five – a team of neural networks that have been trained on a collective 45,000 years of Dota 2 gameplay – just defeated a world champion e-sports team called Team OG in San Francisco last week.
That represents a monumental achievement in designing a system that can upskill itself with a technique called deep reinforcement learning, wherein a bot isn’t taught how to play, but rather how to learn to master the game by figuring out which of its actions bring rewards. The long-term goal is to better understand these systems and apply them to other fields, like robotics.
Special announcement: we’re inviting the entire Internet to play OpenAI Five (whether as a competitor or teammate) at once.
Sign up today! Very excited to see what we learn from observing OpenAI Five in the wild. pic.twitter.com/TaMhxdgVIt
— Greg Brockman (@gdb) April 13, 2019
The Verge reports that the team behind OpenAI Five is wrapping up public demonstrations of the system’s capabilities, and is moving on to building software to allow humans to collaborate with the neural networks in the game. But before that, it’s opening up a platform called Arena to let you take on Five later this week.
TNW Conference 2019 is coming! Check out our glorious new location, inspiring line-up of speakers and activities, and how to be a part of this annual tech extravaganza by clicking here.