Scaling Reinforcement Learning to Robotics by Exploiting the Subsumption Architecture.

Sridhar Mahadevan, Jonathan Connell

Browse the full ICML paper archive.