Robust Reinforcement Learning with Bayesian Optimisation and Quadrature - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Journal Articles Journal of Machine Learning Research Year : 2020

Robust Reinforcement Learning with Bayesian Optimisation and Quadrature


Bayesian optimisation has been successfully applied to a variety of reinforcement learning problems. However, the traditional approach for learning optimal policies in simulators does not utilise the opportunity to improve learning by adjusting certain environment variables: state features that are unobservable and randomly determined by the environment in a physical setting but are controllable in a simulator. This article considers the problem of finding a robust policy while taking into account the impact of environment variables. We present alternating optimisation and quadrature (ALOQ), which uses Bayesian optimisation and Bayesian quadrature to address such settings. We also present transferable ALOQ (TALOQ), for settings where simulator inaccuracies lead to difficulty in transferring the learnt policy to the physical system. We show that our algorithms are robust to the presence of significant rare events, which may not be observable under random sampling but play a substantial role in determining the optimal policy. Experimental results across different domains show that our algorithms learn robust policies efficiently.
Fichier principal
Vignette du fichier
18-216.pdf (1.67 Mo) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-02943567 , version 1 (19-09-2020)


  • HAL Id : hal-02943567 , version 1


Supratik Paul, Konstantinos Chatzilygeroudis, Kamil Ciosek, Jean-Baptiste Mouret, Michael A Osborne, et al.. Robust Reinforcement Learning with Bayesian Optimisation and Quadrature. Journal of Machine Learning Research, 2020, 21, pp.1 - 31. ⟨hal-02943567⟩
87 View
88 Download


Gmail Facebook X LinkedIn More