Behaviour Policy Optimization: Provably Lower Variance Return Estimates for Off-Policy Reinforcement Learning.

Alexander W. Goodall, Edwin Hamel-De le Court, Francesco Belardinelli

Browse the full AAAI paper archive.