Uncertainty Estimation by Flexible Evidential Deep Learning

TL;DR

This paper introduces flexible evidential deep learning ($ ext{F}$-EDL), an advanced uncertainty quantification method that models a more adaptable Dirichlet distribution, improving robustness and reliability in complex scenarios.

Contribution

The paper proposes $ ext{F}$-EDL, extending EDL by predicting a flexible Dirichlet distribution for better uncertainty modeling in diverse and challenging situations.

Findings

$ ext{F}$-EDL achieves state-of-the-art UQ performance.

It generalizes well across classical, long-tailed, and noisy scenarios.

Theoretically, $ ext{F}$-EDL offers advantages over traditional EDL.

Abstract

Uncertainty quantification (UQ) is crucial for deploying machine learning models in high-stakes applications, where overconfident predictions can lead to serious consequences. An effective UQ method must balance computational efficiency with the ability to generalize across diverse scenarios. Evidential deep learning (EDL) achieves efficiency by modeling uncertainty through the prediction of a Dirichlet distribution over class probabilities. However, the restrictive assumption of Dirichlet-distributed class probabilities limits EDL's robustness, particularly in complex or unforeseen situations. To address this, we propose \textit{flexible evidential deep learning} ($\mathcal{F}$-EDL), which extends EDL by predicting a flexible Dirichlet distribution -- a generalization of the Dirichlet distribution -- over class probabilities. This approach provides a more expressive and adaptive representation of uncertainty, significantly enhancing UQ generalization and reliability under challenging scenarios. We theoretically establish several advantages of $\mathcal{F}$-EDL and empirically demonstrate its state-of-the-art UQ performance across diverse evaluation settings, including classical, long-tailed, and noisy in-distribution scenarios.