Robust Adversarial Quantification via Conflict-Aware Evidential Deep Learning

ICLR 2026 Conference Submission13529 Authors

Keywords: Uncertainty Quantification, Adversarial Attack Detection, Evidential Deep Learning

TL;DR: C-EDL boosts robustness in Evidential Deep Learning by detecting conflict from input transformations, improving OOD and adversarial detection without retraining, while keeping high accuracy and low 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 $\approx55\%$) and adversarial data (up to $\approx90\%$), across a range of datasets, attack types, and uncertainty metrics.

Primary Area: probabilistic methods (Bayesian methods, variational inference, sampling, UQ, etc.)

Submission Number: 13529