Calibrated Model-Based Deep Reinforcement Learning

Ali Malik, Volodymyr Kuleshov, Jiaming Song, Danny Nemer, Harlan Seymour, Stefano Ermon
Proceedings of the 36th International Conference on Machine Learning, PMLR 97:4314-4323, 2019.

Abstract

Estimates of predictive uncertainty are important for accurate model-based planning and reinforcement learning. However, predictive uncertainties — especially ones derived from modern deep learning systems — can be inaccurate and impose a bottleneck on performance. This paper explores which uncertainties are needed for model-based reinforcement learning and argues that ideal uncertainties should be calibrated, i.e. their probabilities should match empirical frequencies of predicted events. We describe a simple way to augment any model-based reinforcement learning agent with a calibrated model and show that doing so consistently improves planning, sample complexity, and exploration. On the \textsc{HalfCheetah} MuJoCo task, our system achieves state-of-the-art performance using 50% fewer samples than the current leading approach. Our findings suggest that calibration can improve the performance of model-based reinforcement learning with minimal computational and implementation overhead.

Cite this Paper


BibTeX
@InProceedings{pmlr-v97-malik19a, title = {Calibrated Model-Based Deep Reinforcement Learning}, author = {Malik, Ali and Kuleshov, Volodymyr and Song, Jiaming and Nemer, Danny and Seymour, Harlan and Ermon, Stefano}, booktitle = {Proceedings of the 36th International Conference on Machine Learning}, pages = {4314--4323}, year = {2019}, editor = {Chaudhuri, Kamalika and Salakhutdinov, Ruslan}, volume = {97}, series = {Proceedings of Machine Learning Research}, month = {09--15 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v97/malik19a/malik19a.pdf}, url = {http://proceedings.mlr.press/v97/malik19a.html}, abstract = {Estimates of predictive uncertainty are important for accurate model-based planning and reinforcement learning. However, predictive uncertainties — especially ones derived from modern deep learning systems — can be inaccurate and impose a bottleneck on performance. This paper explores which uncertainties are needed for model-based reinforcement learning and argues that ideal uncertainties should be calibrated, i.e. their probabilities should match empirical frequencies of predicted events. We describe a simple way to augment any model-based reinforcement learning agent with a calibrated model and show that doing so consistently improves planning, sample complexity, and exploration. On the \textsc{HalfCheetah} MuJoCo task, our system achieves state-of-the-art performance using 50% fewer samples than the current leading approach. Our findings suggest that calibration can improve the performance of model-based reinforcement learning with minimal computational and implementation overhead.} }
Endnote
%0 Conference Paper %T Calibrated Model-Based Deep Reinforcement Learning %A Ali Malik %A Volodymyr Kuleshov %A Jiaming Song %A Danny Nemer %A Harlan Seymour %A Stefano Ermon %B Proceedings of the 36th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2019 %E Kamalika Chaudhuri %E Ruslan Salakhutdinov %F pmlr-v97-malik19a %I PMLR %P 4314--4323 %U http://proceedings.mlr.press/v97/malik19a.html %V 97 %X Estimates of predictive uncertainty are important for accurate model-based planning and reinforcement learning. However, predictive uncertainties — especially ones derived from modern deep learning systems — can be inaccurate and impose a bottleneck on performance. This paper explores which uncertainties are needed for model-based reinforcement learning and argues that ideal uncertainties should be calibrated, i.e. their probabilities should match empirical frequencies of predicted events. We describe a simple way to augment any model-based reinforcement learning agent with a calibrated model and show that doing so consistently improves planning, sample complexity, and exploration. On the \textsc{HalfCheetah} MuJoCo task, our system achieves state-of-the-art performance using 50% fewer samples than the current leading approach. Our findings suggest that calibration can improve the performance of model-based reinforcement learning with minimal computational and implementation overhead.
APA
Malik, A., Kuleshov, V., Song, J., Nemer, D., Seymour, H. & Ermon, S.. (2019). Calibrated Model-Based Deep Reinforcement Learning. Proceedings of the 36th International Conference on Machine Learning, in Proceedings of Machine Learning Research 97:4314-4323 Available from http://proceedings.mlr.press/v97/malik19a.html.

Related Material