Google DeepMind: The Podcast cover image

Life is like a game

Google DeepMind: The Podcast

CHAPTER

How to Train a Population of Agents in a Game

Every player acts independently. They only see the own observation. And they're all playing with and against each other. So you have to sort of look around and move through this treedy world from your own first person perspective, but inter act with these other things which see their own first person perspectives. Rather than just creating a single agent, max and his team build an entire class room of them, 30 in total. For each round of the game, he randomly selects a few of the agent from the class to play together on a team. By doing this thousands and thousands of times, each agent will learn from their own experience. But because they are playing with each other too, with their

00:00
Transcript
Play full episode

Remember Everything You Learn from Podcasts

Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
App store bannerPlay store banner