Videos

Multi-Agent Hide and Seek



OpenAI

We’ve observed agents discovering progressively more complex tool use while playing a simple game of hide-and-seek. Through training in our new simulated hide-and-seek environment, agents build a series of six distinct strategies and counterstrategies, some of which we did not know our environment supported. The self-supervised emergent complexity in this simple environment further suggests that multi-agent co-adaptation may one day produce extremely complex and intelligent behavior.

Learn more: https://openai.com/blog/emergent-tool-use/

Source

Similar Posts

24 thoughts on “Multi-Agent Hide and Seek
  1. This can be applied to school, for example, kids who get introduced to school dont know what to do but as they learn they adapt to school they start to listen, stay in order, and make choices to improve. But this process is not really that great because school does not provide much creativity for children. If you think hard enough the "learning" process isn't all that hard. First the student get introduced to new subjects or lessons (math, science, history, and other subjects) after that the student studies the subject then after sometime the test comes then the students regurgitate the information that they learned.

    Anyway I'm bad at English (because I'm 12)

  2. I'm disappointed they didn't put the clip where the agents abused the physics engine to yeet the box out of the map

Comments are closed.

WP2Social Auto Publish Powered By : XYZScripts.com