Reinforcement Learning Can Be More Efficient with Multiple Rewards.

Christoph Dann, Yishay Mansour, Mehryar Mohri

Browse the full ICML paper archive.