Evaluating Adaptation Performance of Hierarchical Deep Reinforcement Learning
Neale Van Stolen, Seung Hyun Kim, Huy T. Tran, Girish Chowdhary
Abstract
Deep Reinforcement Learning has been used to exploit specific environments, but has difficulty transferring learned policies to new situations. This issue poses a problem for practical applications of Reinforcement Learning, as real-world scenarios may introduce unexpected differences that drastically reduce policy performance. We propose the use of differentiated sub-policies governed by a hierarchical controller to support adaptation in such scenarios. We also introduce a confidence- based training process for the hierarchical controller which improves training stability and convergence times. We evaluate these methods in a new Capture the Flag environment designed to explore adaptation in autonomous multi-agent settings.
BibTeX
@inproceedings{icra2020_evaluatingadapta,
title = {Evaluating Adaptation Performance of Hierarchical Deep Reinforcement Learning},
author = {Neale Van Stolen and Seung Hyun Kim and Huy T. Tran and Girish Chowdhary},
booktitle = {ICRA 2020},
year = {2020}
}