Robust Reinforcement Learning via Adversarial training with Langevin Dynamics

We introduce a sampling perspective to tackle the challenging task of training robust Reinforcement Learning (RL) agents. Leveraging the powerful Stochastic Gradient Langevin Dynamics, we present a novel, scalable two-player RL algorithm, which is a sampling variant of the two-player policy gradient method. Our algorithm consistently outperforms existing baselines, in terms of generalization across different training and testing conditions, on several MuJoCo environments. Our experiments also show that, even for objective functions that entirely ignore potential environmental shifts, our sampling approach remains highly robust in comparison to standard RL algorithms.


Year:
Feb 14 2020
Keywords:
Note:
ml-ai
Additional link:
Laboratories:


Note: The status of this file is:


 Record created 2020-02-17, last modified 2020-02-19

PREPRINT:
Download fulltext
PDF

Rate this document:

Rate this document:
1
2
3
 
(Not yet reviewed)