Multi Agent Hide and Seek

Weve observed agents discovering progressively more complex tool use while playing a simple game of hideandseek. Through training in our new simulated hideandseek environment, agents build a series of six distinct strategies and counterstrategies, some of which we did not know our environment supported. The selfsupervised emergent complexity in this simple environment further suggests that multiagent coadaptation may one day produce extremely complex and intelligent behavior. Learn more: op