TibGM: A Transferable and Information-Based Graphical Model Approach for Reinforcement Learning

Adel, T. and Weller, A. (2019) TibGM: A Transferable and Information-Based Graphical Model Approach for Reinforcement Learning. In: 36th International Conference on Machine Learning, ICML 2019, Long Beach, CA, USA, 9-15 June 2019, pp. 71-81.

Full text not currently available from Enlighten.

Publisher's URL: http://proceedings.mlr.press/v97/adel19a.html

Abstract

One of the challenges to reinforcement learning (RL) is scalable transferability among complex tasks. Incorporating a graphical model (GM), along with the rich family of related methods, as a basis for RL frameworks provides potential to address issues such as transferability, generalisation and exploration. Here we propose a flexible GM-based RL framework which leverages efficient inference procedures to enhance generalisation and transfer power. In our proposed transferable and information-based graphical model framework ‘TibGM’, we show the equivalence between our mutual information-based objective in the GM, and an RL consolidated objective consisting of a standard reward maximisation target and a generalisation/transfer objective. In settings where there is a sparse or deceptive reward signal, our TibGM framework is flexible enough to incorporate exploration bonuses depicting intrinsic rewards. We empirically verify improved performance and exploration power.

Item Type:Conference Proceedings
Additional Information:AW acknowledges support from the David MacKay Newton research fellowship at Darwin College and The Alan Turing Institute under EPSRC grant EP/N510129/1 & TU/B/000074.
Status:Published
Refereed:Yes
Glasgow Author(s) Enlighten ID:Hesham, Dr Tameem Adel
Authors: Adel, T., and Weller, A.
College/School:College of Science and Engineering > School of Computing Science
ISSN:2640-3498

University Staff: Request a correction | Enlighten Editors: Update this record