Real-time Uncertainty Decomposition for Online Learning Control

Jonas Umlauft, Armin Lederer, Thomas Beckers, Sandra Hirche

Keywords: uncertainty decomposition, epistemic uncertainty, online learning, real-time control

Abstract:

Safety-critical decisions based on machine learning models require a clear understanding of the involved uncertainties to avoid hazardous or risky situations. While aleatoric uncertainty can be explicitly modeled given a parametric description, epistemic uncertainty rather describes the presents or absence of training data. This paper proposes a novel generic method for modeling
epistemic uncertainty and shows its advantages over existing approaches for neural networks on various data sets. It can be directly combined with aleatoric uncertainty estimates and
allows for prediction in real-time as the inference is sample-free. We exploit this property in a model-based quadcopter control setting and demonstrate how the controller benefits from a differentiation between aleatoric and epistemic uncertainty in online learning of thermal disturbances.

One-sentence Summary: The paper proposes a real-time capable epistemic uncertainty estimation with sample-free inference which allows online learning control applications for which we demonstrate the benefits of aleatoric/epistemic uncertainty decomposition.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Community Implementations: [ 2 code implementations](https://www.catalyzex.com/paper/arxiv:2010.02613/code)

Reviewed Version (pdf): https://openreview.net/references/pdf?id=QjyK4TcFD1