A Deep Reinforcement Learning Approach to Marginalized Importance Sampling with the Successor Representation.

Scott Fujimoto, David Meger, Doina Precup

Browse the full ICML paper archive.