ARL Science Talk video: "Cooperative Multi-Agent Control via Reinforcement Learning"

October 06, 2017

Stanford doctoral candidate and ARL High Performance Computing Research Center (AHPCRC) collaborator, John Mern, discusses how recent developments in deep reinforcement learning enables computers to perform multi-agent cooperative actions through learning in simulated environments. At ARL, cooperative multi-agent control is being combined with extremely low-power computing devices, enabling future capabilities for the US Army in coordinated multi-agent autonomous systems with long-duration intelligent computing capabilities.

 

Last Update / Reviewed: October 6, 2017