Distributional Reinforcement Learning with Dual Expectile-Quantile Regression

arXiv:2305.16877v2 Announce Type: replace
Abstract: Distributional reinforcement learning (RL) has proven useful in multiple benchmarks as it enables approximating the full distribution of returns and makes a better use of environment samples. The commonly used quantile regression approach to distributional RL -- based on asymmetric $L_1$ losses -- provides a flexible and effective way of learning arbitrary return distributions. In practice, it is often improved by using a more efficient, hybrid asymmetric $L_1$-$L_2$ Huber loss for quantile regression. However, by …

abstract arxiv benchmarks cs.ai cs.lg distribution environment losses multiple quantile regression reinforcement reinforcement learning returns samples type