Maximum Entropy-Regularized Multi-Goal Reinforcement Learning

Rui Zhao, Xudong Sun, Volker Tresp
Proceedings of the 36th International Conference on Machine Learning, PMLR 97:7553-7562, 2019.

Abstract

In Multi-Goal Reinforcement Learning, an agent learns to achieve multiple goals with a goal-conditioned policy. During learning, the agent first collects the trajectories into a replay buffer, and later these trajectories are selected randomly for replay. However, the achieved goals in the replay buffer are often biased towards the behavior policies. From a Bayesian perspective, when there is no prior knowledge about the target goal distribution, the agent should learn uniformly from diverse achieved goals. Therefore, we first propose a novel multi-goal RL objective based on weighted entropy. This objective encourages the agent to maximize the expected return, as well as to achieve more diverse goals. Secondly, we developed a maximum entropy-based prioritization framework to optimize the proposed objective. For evaluation of this framework, we combine it with Deep Deterministic Policy Gradient, both with or without Hindsight Experience Replay. On a set of multi-goal robotic tasks of OpenAI Gym, we compare our method with other baselines and show promising improvements in both performance and sample-efficiency.

Cite this Paper


BibTeX
@InProceedings{pmlr-v97-zhao19d, title = {Maximum Entropy-Regularized Multi-Goal Reinforcement Learning}, author = {Zhao, Rui and Sun, Xudong and Tresp, Volker}, booktitle = {Proceedings of the 36th International Conference on Machine Learning}, pages = {7553--7562}, year = {2019}, editor = {Chaudhuri, Kamalika and Salakhutdinov, Ruslan}, volume = {97}, series = {Proceedings of Machine Learning Research}, month = {09--15 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v97/zhao19d/zhao19d.pdf}, url = {http://proceedings.mlr.press/v97/zhao19d.html}, abstract = {In Multi-Goal Reinforcement Learning, an agent learns to achieve multiple goals with a goal-conditioned policy. During learning, the agent first collects the trajectories into a replay buffer, and later these trajectories are selected randomly for replay. However, the achieved goals in the replay buffer are often biased towards the behavior policies. From a Bayesian perspective, when there is no prior knowledge about the target goal distribution, the agent should learn uniformly from diverse achieved goals. Therefore, we first propose a novel multi-goal RL objective based on weighted entropy. This objective encourages the agent to maximize the expected return, as well as to achieve more diverse goals. Secondly, we developed a maximum entropy-based prioritization framework to optimize the proposed objective. For evaluation of this framework, we combine it with Deep Deterministic Policy Gradient, both with or without Hindsight Experience Replay. On a set of multi-goal robotic tasks of OpenAI Gym, we compare our method with other baselines and show promising improvements in both performance and sample-efficiency.} }
Endnote
%0 Conference Paper %T Maximum Entropy-Regularized Multi-Goal Reinforcement Learning %A Rui Zhao %A Xudong Sun %A Volker Tresp %B Proceedings of the 36th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2019 %E Kamalika Chaudhuri %E Ruslan Salakhutdinov %F pmlr-v97-zhao19d %I PMLR %P 7553--7562 %U http://proceedings.mlr.press/v97/zhao19d.html %V 97 %X In Multi-Goal Reinforcement Learning, an agent learns to achieve multiple goals with a goal-conditioned policy. During learning, the agent first collects the trajectories into a replay buffer, and later these trajectories are selected randomly for replay. However, the achieved goals in the replay buffer are often biased towards the behavior policies. From a Bayesian perspective, when there is no prior knowledge about the target goal distribution, the agent should learn uniformly from diverse achieved goals. Therefore, we first propose a novel multi-goal RL objective based on weighted entropy. This objective encourages the agent to maximize the expected return, as well as to achieve more diverse goals. Secondly, we developed a maximum entropy-based prioritization framework to optimize the proposed objective. For evaluation of this framework, we combine it with Deep Deterministic Policy Gradient, both with or without Hindsight Experience Replay. On a set of multi-goal robotic tasks of OpenAI Gym, we compare our method with other baselines and show promising improvements in both performance and sample-efficiency.
APA
Zhao, R., Sun, X. & Tresp, V.. (2019). Maximum Entropy-Regularized Multi-Goal Reinforcement Learning. Proceedings of the 36th International Conference on Machine Learning, in Proceedings of Machine Learning Research 97:7553-7562 Available from http://proceedings.mlr.press/v97/zhao19d.html.

Related Material