Data to 14 Oct 2026

On the Connection Between Adversarial Robustness and Saliency Map Interpretability

arXiv 1905.04172 ↗First posted 10 May 2019stat.ML cs.CV cs.LGcited by 80

Not on the map. Why?

Abstract (arXiv)

Recent studies on the adversarial vulnerability of neural networks have shown that models trained to be more robust to adversarial attacks exhibit more interpretable saliency maps than their non-robust counterparts. We aim to quantify this behavior by considering the alignment between input image and saliency map. We hypothesize that as the distance to the decision boundary grows,so does the alignment. This connection is strictly true in the case of linear models. We confirm these theoretical findings with experiments based on models trained with a local Lipschitz regularization and identify where the non-linear nature of neural networks weakens the relation.

Key facts. Cited by 80 papers in the map (data through 14 Oct 2026). First posted on arXiv 10 May 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 19; 2022: 20; 2023: 15; 2024: 7; 2025: 11; 2026: 8.

80 papers.

Sort:

By area:

2022From Anecdotal Evidence to Quantitative Evaluation Methods: A Systematic Review on Evaluating Explainable AIcited by 302
2022Toward Transparent AI: A Survey on Interpreting the Inner Structures of Deep Neural Networkscited by 129
2021Interpretable Deep Learning: Interpretation, Interpretability, Trustworthiness, and Beyondcited by 118
2022Robustness and Accuracy Could Be Reconcilable by (Proper) Definitioncited by 111
2022Diffusion Visual Counterfactual Explanationscited by 96
2021Trustworthy AI: A Computational Perspectivecited by 93
2021Robust Explainability: A Tutorial on Gradient-Based Attribution Methods for Deep Neural Networkscited by 48
2023Adversarial Counterfactual Visual Explanationscited by 47
2022Spurious Features Everywhere -- Large-Scale Detection of Harmful Spurious Features in ImageNetcited by 41
2023Fast Diffusion-Based Counterfactuals for Shortcut Removal and Generationcited by 41
2021Leave-one-out Unfairnesscited by 36
2022A Comprehensive Survey on Trustworthy Recommender Systemscited by 32

38 more

2022Sparse Visual Counterfactual Explanations in Image Spacecited by 30
2023Interpretability Benchmark for Evaluating Spatial Misalignment of Prototypical Parts Explanationscited by 27
2022What Makes a Good Explanation?: A Harmonized View of Properties of Explanationscited by 22
2021Robust Models Are More Interpretable Because Attributions Look Normalcited by 21
2021Attack to Fool and Explain Deep Networkscited by 17
2022Do Perceptually Aligned Gradients Imply Adversarial Robustness?cited by 15
2021Identifying Untrustworthy Predictions in Neural Networks by Geometric Gradient Analysiscited by 13
2021Adversarial Visual Robustness by Causal Interventioncited by 13
2022Fooling Explanations in Text Classifierscited by 13
2023CLIPAG: Towards Generator-Free Text-to-Image Generationcited by 12
2024An Unsupervised Approach to Achieve Supervised-Level Explainability in Healthcare Recordscited by 12
2021Unifying Model Explainability and Robustness for Joint Text Classification and Rationale Extractioncited by 11
2022How Does Frequency Bias Affect the Robustness of Neural Image Classifiers against Common Corruption and Adversarial Perturbations?cited by 11
2021When and How to Fool Explainable Models (and Humans) with Adversarial Examplescited by 10
2024Concept Alignmentcited by 10
2022How explainable are adversarially-robust CNNs?cited by 9
2023Bayesian Neural Networks Avoid Encoding Complex and Perturbation-Sensitive Conceptscited by 9
2023ACAT: Adversarial Counterfactual Attention for Classification and Detection in Medical Imagingcited by 9
2021Impact of Spatial Frequency Based Constraints on Adversarial Robustnesscited by 8
2021Now You See It, Now You Dont: Adversarial Vulnerabilities in Computational Pathologycited by 7
2022A Survey on AI Sustainability: Emerging Trends on Learning Algorithms and Research Challengescited by 7
2022SoK: Modeling Explainability in Security Analytics for Interpretability, Trustworthiness, and Usabilitycited by 7
2025Superposition as Lossy Compression: Measure with Sparse Autoencoders and Connect to Adversarial Vulnerabilitycited by 7
2022Visualizing Automatic Speech Recognition -- Means for a Better Understanding?cited by 6
2022Adversarial Parameter Attack on Deep Neural Networkscited by 6
2022On the explainable properties of 1-Lipschitz Neural Networks: An Optimal Transport Perspectivecited by 6
2023Interpretability is a Kind of Safety: An Interpreter-based Ensemble for Adversary Defensecited by 6
2022On the Robustness of Explanations of Deep Neural Network Models: A Surveycited by 5
2023Adversarial Attacks and Defenses in Machine Learning-Powered Networks: A Contemporary Surveycited by 5
2023Class-Conditioned Transformation for Enhanced Robust Image Classificationcited by 5
2023SAFE: Saliency-Aware Counterfactual Explanations for DNN-based Automated Driving Systemscited by 5
2023Explainable artificial intelligence approaches for brain-computer interfaces: a review and design spacecited by 5
2024Improving the Evaluation and Actionability of Explanation Methods for Multivariate Time Series Classificationcited by 5
2021Investigating the significance of adversarial attacks and their relation to interpretability for radar-based human activity recognition systemscited by 4
2021Improving the trustworthiness of image classification models by utilizing bounding-box annotationscited by 4
2021Get Fooled for the Right Reason: Improving Adversarial Robustness through a Teacher-guided Curriculum Learning Approachcited by 4
2022Robustness of Explanation Methods for NLP Modelscited by 3
2025Adversarially Pretrained Transformers May Be Universally Robust In-Context Learnerscited by 3

50 most cited of 80 All 80 (JSON)