Boosting Reinforcement Learning with Strongly Delayed Feedback Through Auxiliary Short Delays

This paper presents Auxiliary-Delayed Reinforcement Learning (AD-RL), a novel framework that addresses the challenge of learning with long observation delays in reinforcement learning. The key innovation is leveraging an auxiliary short-delayed task to accelerate learning on the primary long-delayed task, maintaining performance quality in stochastic environments.

Authors: Qingyuan Wu, Simon Sinong Zhan, Yixuan Wang, Yuhui Wang, Chung-Wei Lin, Chen Lv, Qi Zhu, Jürgen Schmidhuber, Chao Huang

@inproceedings{wu2024boosting, title={Boosting Reinforcement Learning with Strongly Delayed Feedback Through Auxiliary Short Delays}, author={Wu, Qingyuan and Zhan, Sinong Simon and Wang, Yixuan and Wang, Yuhui and Lin, Chung-Wei and Lv, Chen and Zhu, Qi and Schmidhuber, Jürgen and Huang, Chao}, booktitle={International Conference on Machine Learning (ICML)}, year={2024}, url={https://arxiv.org/abs/2402.03141} }