Uncertainty-Based Out-of-Distribution Classification in Deep Reinforcement Learning

by   Andreas Sedlmeier, et al.
Universität München

Robustness to out-of-distribution (OOD) data is an important goal in building reliable machine learning systems. Especially in autonomous systems, wrong predictions for OOD inputs can cause safety critical situations. As a first step towards a solution, we consider the problem of detecting such data in a value-based deep reinforcement learning (RL) setting. Modelling this problem as a one-class classification problem, we propose a framework for uncertainty-based OOD classification: UBOOD. It is based on the effect that an agent's epistemic uncertainty is reduced for situations encountered during training (in-distribution), and thus lower than for unencountered (OOD) situations. Being agnostic towards the approach used for estimating epistemic uncertainty, combinations with different uncertainty estimation methods, e.g. approximate Bayesian inference methods or ensembling techniques are possible. We further present a first viable solution for calculating a dynamic classification threshold, based on the uncertainty distribution of the training data. Evaluation shows that the framework produces reliable classification results when combined with ensemble-based estimators, while the combination with concrete dropout-based estimators fails to reliably detect OOD situations. In summary, UBOOD presents a viable approach for OOD classification in deep RL settings by leveraging the epistemic uncertainty of the agent's value function.


Uncertainty-Based Out-of-Distribution Detection in Deep Reinforcement Learning

We consider the problem of detecting out-of-distribution (OOD) samples i...

Policy Entropy for Out-of-Distribution Classification

One critical prerequisite for the deployment of reinforcement learning s...

Disentangling Epistemic and Aleatoric Uncertainty in Reinforcement Learning

Characterizing aleatoric and epistemic uncertainty on the predicted rewa...

MACEst: The reliable and trustworthy Model Agnostic Confidence Estimator

Reliable Confidence Estimates are hugely important for any machine learn...

Estimating Risk and Uncertainty in Deep Reinforcement Learning

This paper demonstrates a novel method for separately estimating aleator...

Multivariate Deep Evidential Regression

There is significant need for principled uncertainty reasoning in machin...

Quantifying Multimodality in World Models

Model-based Deep Reinforcement Learning (RL) assumes the availability of...

Code Repositories


Cross-domain Robot Navigation with Deep Reinforcement Learning

view repo


Modifiable version of the LunarLander gym environment

view repo

Please sign up or login with your details

Forgot password? Click here to reset