
Multi-Agent Hide and Seek | (dot)AI

(dot)AI | A space by MATTIA FALDUTI
Humans evolve through generations. Today also virtual agents do it.
One year ago, OpenAI described how the use of standard reinforcement algorithms permits a multi-agent competition of hide-and-seek.
Spoiler Alert: the teams discover themselves unpredictable strategies.
Below, the abstract of the paper written by the OpenAI team.
