Mind the Uncertainty: Risk-Aware and Actively Exploring Model-Based Reinforcement Learning

Marin Vlastelica, Sebastian Blaes, Cristina Pinneri, Georg Martius

Keywords: CEM, data-driven MPC, uncertainty, model-based RL

TL;DR: Model-based method with an ensemble of probabilistic models used for managing aleatoric/ epistemic uncertainty and probabilistic safety constraints.

Abstract: We introduce a simple but effective method for managing risk in model-based reinforcement learning with trajectory sampling that involves probabilistic safety constraints and balancing of optimism in the face of epistemic uncertainty and pessimism in the face of aleatoric uncertainty of an ensemble of stochastic neural networks. Various experiments indicate that the separation of uncertainties is essential to performing well with data-driven MPC approaches in uncertain and safety-critical control environments.

Already Accepted Paper At Another Venue: already accepted somewhere else