Practical Probabilistic Model-based Deep Reinforcement Learning by Integrating Dropout Uncertainty and Trajectory Sampling

TL;DR

This paper introduces DPETS, a novel probabilistic model-based reinforcement learning method that combines dropout uncertainty and trajectory sampling to improve prediction stability, accuracy, and control in complex tasks.

Contribution

The paper proposes a new dropout-based probabilistic ensemble with trajectory sampling (DPETS) that enhances uncertainty prediction and control in model-based RL, outperforming existing methods.

Findings

DPETS achieves higher average returns on Mujoco benchmarks.

DPETS converges faster than related MBRL approaches.

DPETS outperforms well-known model-free baselines in sample efficiency.

Abstract

This paper addresses the prediction stability, prediction accuracy and control capability of the current probabilistic model-based reinforcement learning (MBRL) built on neural networks. A novel approach dropout-based probabilistic ensembles with trajectory sampling (DPETS) is proposed where the system uncertainty is stably predicted by combining the Monte-Carlo dropout and trajectory sampling in one framework. Its loss function is designed to correct the fitting error of neural networks for more accurate prediction of probabilistic models. The state propagation in its policy is extended to filter the aleatoric uncertainty for superior control capability. Evaluated by several Mujoco benchmark control tasks under additional disturbances and one practical robot arm manipulation task, DPETS outperforms related MBRL approaches in both average return and convergence velocity while achieving superior performance than well-known model-free baselines with significant sample efficiency. The open source code of DPETS is available at https://github.com/mrjun123/DPETS.