Robust Adversarial Quantification via Conflict-Aware Evidential Deep Learning

2025/06/06 by Charmaine Barker, Barker, Charmaine, Daniel Bethell +3 · 1 citation
Computer Science · #Advanced Neural Network Applications #Adversarial Robustness in Machine Learning #Artificial Intelligence (cs.AI) #Explainable Artificial Intelligence (XAI) #FOS: Computer and information sciences #Machine Learning (cs.LG)

paper · doi:10.48550/arxiv.2506.05937

openalex publication_date 2025/06/06 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/29

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 ≈55%) and adversarial data (up to ≈90%), across a range of datasets, attack types, and uncertainty metrics.

Citations

• Uncertainty Quantification for Deep Regression using Contextualised Normalizing Flows
• Guided Uncertainty Learning Using a Post-Hoc Evidential Meta-Model
• Uncertainty Estimation by Density Aware Evidential Deep Learning
• Reliable Conflictive Multi-View Learning
• Make Me a BNN: A Simple Strategy for Estimating Bayesian Uncertainty from Pre-trained Models
• Robust Uncertainty Quantification Using Conformalised Monte Carlo Prediction
• Quantification of Uncertainty with Adversarial Models
• Learn to Accumulate Evidence from All Training Samples: Theory and Practice
• GPT-4 Technical Report
• Uncertainty Estimation by Fisher Information-based Evidential Deep Learning
• Post-hoc Uncertainty Learning using a Dirichlet Meta-Model
• OpenOOD: Benchmarking Generalized Out-of-Distribution Detection
• Priors in Bayesian Deep Learning: A Review
• A review of uncertainty quantification in deep learning: Techniques, applications and challenges
• Evaluating Robustness of Predictive Uncertainty Estimation: Are Dirichlet-based Models Reliable?
• An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
• Posterior Network: Uncertainty Estimation without OOD Samples via Density-Based Pseudo-Counts
• Simple and Principled Uncertainty Estimation with Deterministic Deep Learning via Distance Awareness
• Evidential Deep Learning to Quantify Classification Uncertainty
• Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
• On Calibration of Modern Neural Networks
• What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?
• Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles
• Variational Inference: A Review for Statisticians
• Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning
• Lower Upper Bound Estimation Method for Construction of Neural Network-Based Prediction Intervals
• A Survey on Uncertainty Quantification Methods for Deep Learning
• Gradient-based learning applied to document recognition

Cited by

• AEGIS: A Semantic GAN and Evidential Learning Framework for Robust Adversarial Detection in Vision Sensors

Related

• Are Uncertainty Quantification Capabilities of Evidential Deep Learning a Mirage?
• A Comprehensive Survey on Evidential Deep Learning and Its Applications
• Uncertainty Estimation by Flexible Evidential Deep Learning
• Knowledge from Uncertainty in Evidential Deep Learning
• Calibrating LLMs with Information-Theoretic Evidential Deep Learning