Self-Paced Context Evaluation for Contextual Reinforcement Learning

authored by

Theresa Eimer, André Biedenkapp, Frank Hutter, Marius Lindauer

Abstract

Reinforcement learning (RL) has made a lot of advances for solving a single problem in a given environment; but learning policies that generalize to unseen variations of a problem remains challenging. To improve sample efficiency for learning on such instances of a problem domain, we present Self-Paced Context Evaluation (SPaCE). Based on self-paced learning, \spc automatically generates \task curricula online with little computational overhead. To this end, SPaCE leverages information contained in state values during training to accelerate and improve training performance as well as generalization capabilities to new instances from the same problem domain. Nevertheless, SPaCE is independent of the problem domain at hand and can be applied on top of any RL agent with state-value function approximation. We demonstrate SPaCE's ability to speed up learning of different value-based RL agents on two environments, showing better generalization capabilities and up to 10x faster learning compared to naive approaches such as round robin or SPDRL, as the closest state-of-the-art approach.

Organisation(s)

Institute of Information Processing
Machine Learning Section

External Organisation(s)

University of Freiburg
Bosch Center for Artificial Intelligence (BCAI)

Type

Conference contribution

No. of pages

14

Publication date

2021

Publication status

E-pub ahead of print

Peer reviewed

Yes

Electronic version(s)

https://www.tnt.uni-hannover.de/papers/data/1454/space.pdf (Access: Open)
https://arxiv.org/abs/2106.05110 (Access: Open)
http://proceedings.mlr.press/v139/eimer21a/eimer21a.pdf (Access: Open)