mlsanityData to 7 Oct 2026

Robust Ranking Explanations

arXiv 2307.04024 ↗First posted 8 Jul 2023cs.LG cs.CRcited by 0

Not on the map. Why?

Abstract (arXiv)

Robust explanations of machine learning models are critical to establish human trust in the models. Due to limited cognition capability, most humans can only interpret the top few salient features. It is critical to make top salient features robust to adversarial attacks, especially those against the more vulnerable gradient-based explanations. Existing defense measures robustness using ℓp-norms, which have weaker protection power. We define explanation thickness for measuring salient features ranking stability, and derive tractable surrogate bounds of the thickness to design the textitR2ET algorithm to efficiently maximize the thickness and anchor top salient features. Theoretically, we prove a connection between R2ET and adversarial training. Experiments with a wide spectrum of network architectures and data modalities, including brain networks, demonstrate that R2ET attains higher explanation robustness under stealthy attacks while retaining accuracy.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 63 of its 93 references that are in the map. First posted on arXiv 8 Jul 2023.

Builds on

63 of 93 references.

References · 63

  1. 2016Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localizationcited by 6,690
  2. 2017A Unified Approach to Interpreting Model Predictionscited by 6,488
  3. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,326
  4. 2016"Why Should I Trust You?": Explaining the Predictions of Any Classifiercited by 5,040
  5. 2017Axiomatic Attribution for Deep Networkscited by 3,350
  6. 2015Learning Deep Features for Discriminative Localizationcited by 2,430
  7. 2019Theoretically Principled Trade-off between Robustness and Accuracycited by 1,510
  8. 2017Learning Important Features Through Propagating Activation Differencescited by 1,454
  9. 2018Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examplescited by 1,259
  10. 2018Sanity Checks for Saliency Mapscited by 1,149
  11. 2017SmoothGrad: removing noise by adding noisecited by 1,097
  12. 2018Robustness May Be at Odds with Accuracycited by 873
38 more
  1. 2020Fast is better than free: Revisiting adversarial trainingcited by 723
  2. 2019Explaining Machine Learning Classifiers through Diverse Counterfactual Explanationscited by 612
  3. 2019ERASER: A Benchmark to Evaluate Rationalized NLP Modelscited by 523
  4. 2018Towards Robust Interpretability with Self-Explaining Neural Networkscited by 505
  5. 2017Interpretation of Neural Networks is Fragilecited by 443
  6. 2019Is Attention Interpretable?cited by 387
  7. 2016European Union regulations on algorithmic decision-making and a "right to explanation"cited by 343
  8. 2018On the Robustness of Interpretability Methodscited by 310
  9. 2019On Gradient Descent Ascent for Nonconvex-Concave Minimax Problemscited by 266
  10. 2018Explanations based on the Missing: Towards Contrastive Explanations with Pertinent Negativescited by 261
  11. 2019Adversarial Attacks and Defenses in Images, Graphs and Text: A Reviewcited by 234
  12. 2019Explanations can be manipulated and geometry is to blamecited by 215
  13. 2020GraphLIME: Local Interpretable Model Explanations for Graph Neural Networkscited by 209
  14. 2017Formal Guarantees on the Robustness of a Classifier against Adversarial Manipulationcited by 182
  15. 2018Robustness via curvature regularization, and vice versacited by 174
  16. 2019Explainability Techniques for Graph Convolutional Networkscited by 160
  17. 2020Understanding and Mitigating the Tradeoff Between Robustness and Accuracycited by 148
  18. 2021On the Convergence and Robustness of Adversarial Trainingcited by 127
  19. 2019Explaining Deep Neural Networks with a Polynomial Time Algorithm for Shapley Values Approximationcited by 115
  20. 2019Fooling Neural Network Interpretations via Adversarial Model Manipulationcited by 112
  21. 2021Towards Robust and Reliable Algorithmic Recoursecited by 101
  22. 2019Neural Network Attributions: A Causal Perspectivecited by 77
  23. 2015Learning with a Strong Adversarycited by 68
  24. 2018Theoretical Impediments to Machine Learning With Seven Sparks from the Causal Revolutioncited by 66
  25. 2021Robust Explainability: A Tutorial on Gradient-Based Attribution Methods for Deep Neural Networkscited by 48
  26. 2019Robust Attribution Regularizationcited by 46
  27. 2020Towards Robust Explanations for Deep Neural Networkscited by 43
  28. 2020Transferable, Controllable, and Inconspicuous Adversarial Attacks on Person Re-identification With Deep Mis-Rankingcited by 42
  29. 2020Smoothed Geometry for Robust Attributioncited by 38
  30. 2020Towards Understanding Fast Adversarial Trainingcited by 32
  31. 2018Theoretical Analysis of Adversarial Learning: A Minimax Approachcited by 31
  32. 2020Augmented Lagrangian Adversarial Attackscited by 28
  33. 2019Targeted Mismatch Adversarial Attack: Query with a Flower to Retrieve the Towercited by 26
  34. 2020Boundary thickness and robustness in learning modelscited by 26
  35. 2020Improving Adversarial Robustness via Unlabeled Out-of-Domain Datacited by 23
  36. 2020A simple defense against adversarial attacks on heatmap explanationscited by 23
  37. 2022Robust Explanation Constraints for Neural Networkscited by 23
  38. 2021QAIR: Practical Query-efficient Black-Box Attacks for Image Retrievalcited by 22

50 most cited of 63 All 63 (JSON)

Built on by

Not cited yet.