Mind the Uncertainty: Risk-Aware and Actively Exploring Model-Based Reinforcement Learning
Machine Learning
2023-09-12 v1 Artificial Intelligence
Robotics
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.
Cite
@article{arxiv.2309.05582,
title = {Mind the Uncertainty: Risk-Aware and Actively Exploring Model-Based Reinforcement Learning},
author = {Marin Vlastelica and Sebastian Blaes and Cristina Pineri and Georg Martius},
journal= {arXiv preprint arXiv:2309.05582},
year = {2023}
}