In multi-task reinforcement learning there are two main challenges: at training time, the ability to learn different policies with a single model; at test time, inferring which of those policies applying without an external signal. In the case of continual reinforcement learning a third challenge arises: learning tasks sequentially without forgetting the previous ones. In this paper, we tackle these challenges by proposing DisCoRL, an approach combining state representation learning and policy distillation. We experiment on a sequence of three simulated 2D navigation tasks with a 3 wheel omni-directional robot. Moreover, we tested our approach's robustness by transferring the final policy into a real life setting. The policy can solve all tasks and automatically infer which one to run.
@article{arxiv.1907.05855,
title = {DisCoRL: Continual Reinforcement Learning via Policy Distillation},
author = {René Traoré and Hugo Caselles-Dupré and Timothée Lesort and Te Sun and Guanghang Cai and Natalia Díaz-Rodríguez and David Filliat},
journal= {arXiv preprint arXiv:1907.05855},
year = {2019}
}
Comments
arXiv admin note: text overlap with arXiv:1906.04452