On a connection between importance sampling and the likelihood ratio policy gradient

Likelihood ratio policy gradient methods have been some of the most successful reinforcement learning algorithms, especially for learning on physical systems. We describe how the likelihood ratio policy gradient can be derived from an im-portance sampling perspective. This derivation highlights how likelihood ratio methods under-use past experience by (i) using the past experience to estimate only the gradient of the expected return U(θ) at the current policy parameteri-zation θ, rather than to obtain a more complete estimate of U(θ), and (ii) using past experience under the current policy only rather than using all past experience to improve the estimates. We present a new policy search method, which lever-ages both of these observations as well as generalized baselines—a new technique which generalizes commonly used baseline techniques for policy gradient meth-ods. Our algorithm outperforms standard likelihood ratio policy gradient algo-rithms on several testbeds.