Objective Mismatch in Model-based Reinforcement Learning

Learning for Dynamics & Control (L4DC)


Model-based reinforcement learning (MBRL) is a powerful framework for data-efficiently learning control of continuous tasks. Recent work in MBRL has mostly focused on using more advanced function approximators and planning schemes, with little development of the general framework. In this paper, we identify a fundamental issue of the standard MBRL framework – what we call objective mismatch. Objective mismatch arises when one objective is optimized in the hope that a second, often uncorrelated, metric will also be optimized. In the context of MBRL, we characterize the objective mismatch between training the forward dynamics model w.r.t. the likelihood of the one-step ahead prediction, and the overall goal of improving performance on a downstream control task. For example, this issue can emerge with the realization that dynamics models effective for a specific task do not necessarily need to be globally accurate, and vice versa globally accurate models might not be sufficiently accurate locally to obtain good control performance on a specific task. In our experiments, we study this objective mismatch issue and demonstrate that the likelihood of one-step ahead predictions is not always correlated with control performance. This observation highlights a critical limitation in the MBRL framework which will require further research to be fully understood and addressed. We propose an initial method to mitigate the mismatch issue by re-weighting dynamics model training. Building on it, we conclude with a discussion about other potential directions of research for addressing this issue.

Related Publications

All Publications

AISTATS - April 13, 2021

Aligning Time Series on Incomparable Spaces

Samuel Cohen, Giulia Luise, Alexander Terenin, Brandon Amos, Marc Peter Deisenroth

SIGGRAPH - August 9, 2021

Mixture of Volumetric Primitives for Efficient Neural Rendering

Stephen Lombardi, Tomas Simon, Gabriel Schwartz, Michael Zollhoefer, Yaser Sheikh, Jason Saragih

AISTATS - April 13, 2021

Continual Learning using a Bayesian Nonparametric Dictionary of Weight Factors

Nikhil Mehta, Kevin J Liang, Vinay K Verma, Lawrence Carin

NeurIPS - December 6, 2020

Improved Sample Complexity for Incremental Autonomous Exploration in MDPs

Jean Tarbouriech, Matteo Pirotta, Michal Valko, Alessandro Lazaric

To help personalize content, tailor and measure ads, and provide a safer experience, we use cookies. By clicking or navigating the site, you agree to allow our collection of information on and off Facebook through cookies. Learn more, including about available controls: Cookies Policy