Robust Adversarial Quantification via Conflict-Aware Evidential Deep Learning

TL;DR

This paper introduces C-EDL, a lightweight post-hoc method that enhances the robustness of evidential deep learning models against adversarial and out-of-distribution inputs by leveraging conflict-aware uncertainty calibration.

Contribution

C-EDL is a novel, post-hoc approach that improves adversarial and OOD robustness of EDL models without retraining, using conflict-aware transformations and disagreement quantification.

Findings

C-EDL significantly reduces coverage for OOD data by up to 55%.

C-EDL achieves up to 90% reduction in coverage for adversarial data.

C-EDL maintains high in-distribution accuracy with low computational overhead.

Abstract

Reliability of deep learning models is critical for deployment in high-stakes applications, where out-of-distribution or adversarial inputs may lead to detrimental outcomes. Evidential Deep Learning, an efficient paradigm for uncertainty quantification, models predictions as Dirichlet distributions of a single forward pass. However, EDL is particularly vulnerable to adversarially perturbed inputs, making overconfident errors. Conflict-aware Evidential Deep Learning (C-EDL) is a lightweight post-hoc uncertainty quantification approach that mitigates these issues, enhancing adversarial and OOD robustness without retraining. C-EDL generates diverse, task-preserving transformations per input and quantifies representational disagreement to calibrate uncertainty estimates when needed. C-EDL's conflict-aware prediction adjustment improves detection of OOD and adversarial inputs, maintaining high in-distribution accuracy and low computational overhead. Our experimental evaluation shows that C-EDL significantly outperforms state-of-the-art EDL variants and competitive baselines, achieving substantial reductions in coverage for OOD data (up to $\approx$55%) and adversarial data (up to $\approx$90%), across a range of datasets, attack types, and uncertainty metrics.