mlsanityData to 14 Oct 2026

Towards Certification of Uncertainty Calibration under Adversarial Attacks

arXiv 2405.13922 ↗First posted 22 May 2024cs.LG stat.MLcited by 5

Not on the map. Why?

Abstract (arXiv)

Since neural classifiers are known to be sensitive to adversarial perturbations that alter their accuracy, textitcertification methods have been developed to provide provable guarantees on the insensitivity of their predictions to such perturbations. Furthermore, in safety-critical applications, the frequentist interpretation of the confidence of a classifier (also known as model calibration) can be of utmost importance. This property can be measured via the Brier score or the expected calibration error. We show that attacks can significantly harm calibration, and thus propose certified calibration as worst-case bounds on calibration under adversarial perturbations. Specifically, we produce analytic bounds for the Brier score and approximate bounds via the solution of a mixed-integer program on the expected calibration error. Finally, we propose novel calibration attacks and demonstrate how they can improve model calibration through textitadversarial calibration training.

Key facts. Cited by 5 papers in the map (data through 14 Oct 2026). Builds on 44 of its 68 references that are in the map. First posted on arXiv 22 May 2024.

Builds on

44 of 68 references.

References · 44

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2014Explaining and Harnessing Adversarial Examplescited by 7,364
  3. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
  4. 2015Rethinking the Inception Architecture for Computer Visioncited by 5,475
  5. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,183
  6. 2013Intriguing properties of neural networkscited by 5,155
  7. 2017On Calibration of Modern Neural Networkscited by 4,960
  8. 2017Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithmscited by 4,349
  9. 2016Identity Mappings in Deep Residual Networkscited by 1,965
  10. 2019Certified Adversarial Robustness via Randomized Smoothingcited by 1,340
  11. 2016Adversarial Machine Learning at Scalecited by 936
  12. 2020A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challengescited by 819
32 more
  1. 2018Evidential Deep Learning to Quantify Classification Uncertaintycited by 732
  2. 2017Provable defenses against adversarial examples via the convex outer adversarial polytopecited by 567
  3. 2018Threat of Adversarial Attacks on Deep Learning in Computer Vision: A Surveycited by 460
  4. 2019Measuring Calibration in Deep Learningcited by 446
  5. 2018Efficient Neural Network Robustness Certification with General Activation Functionscited by 445
  6. 2020Calibrating Deep Neural Networks using Focal Losscited by 357
  7. 2019Provably Robust Deep Learning via Adversarially Trained Smoothed Classifierscited by 330
  8. 2018On the Effectiveness of Interval Bound Propagation for Training Verifiably Robust Modelscited by 328
  9. 2019Verified Uncertainty Calibrationcited by 309
  10. 2017Formal Verification of Piece-Wise Linear Feed-Forward Neural Networkscited by 242
  11. 2019Algorithms for Verifying Deep Neural Networkscited by 221
  12. 2018Efficient Formal Safety Analysis of Neural Networkscited by 165
  13. 2020Evaluation of Neural Architectures Trained with Square Loss vs Cross-Entropy in Classification Taskscited by 138
  14. 2020MACER: Attack-free and Scalable Robust Training via Maximizing Certified Radiuscited by 114
  15. 2015Global Convergence of ADMM in Nonconvex Nonsmooth Optimizationcited by 98
  16. 2021Soft Calibration Objectives for Neural Networkscited by 92
  17. 2015Posterior calibration and exploratory analysis for natural language processing modelscited by 76
  18. 2020Opportunities and Challenges in Deep Learning Adversarial Robustness: A Surveycited by 72
  19. 2019Non-Parametric Calibration for Classificationcited by 57
  20. 2022Certified Training: Small Boxes are All You Needcited by 52
  21. 2021SmoothMix: Training Confidence-calibrated Smoothed Classifiers for Certified Robustnesscited by 49
  22. 2022Adversarial Robustness of Deep Neural Networks: A Survey from a Formal Verification Perspectivecited by 42
  23. 2020Evaluating Robustness of Predictive Uncertainty Estimation: Are Dirichlet-based Models Reliable?cited by 39
  24. 2020Stronger Calibration Lower Bounds via Sidesteppingcited by 38
  25. 2020Certifying Confidence via Randomized Smoothingcited by 36
  26. 2018Cautious Deep Learningcited by 35
  27. 2020Towards Trustworthy Predictions from Deep Neural Networks with Fast Adversarial Calibrationcited by 34
  28. 2021Meta-Calibration: Learning of Model Calibration Using Differentiable Expected Calibration Errorcited by 29
  29. 2020Improving Calibration through the Relationship with Adversarial Robustnesscited by 26
  30. 2022Robust Models are less Over-Confidentcited by 25
  31. 2021Disrupting Deep Uncertainty Estimation Without Harming Accuracycited by 19
  32. 2019Constrained Clustering: General Pairwise and Cardinality Constraintscited by 5

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 2; 2026: 3.

5 papers.

  1. 2025HQFNN: A Compact Quantum-Fuzzy Neural Network for Accurate Image Classificationcited by 0
  2. 2025Weak Links in LinkedIn: Enhancing Fake Profile Detection in the Age of LLMscited by 0
  3. 2026Reinforcement Learning Disrupts Gradient-Based Adversarial Optimizationcited by 0
  4. 2026AEGIS: A Semantic GAN and Evidential Learning Frameworkfor Robust Adversarial Detection in Vision Sensorscited by 0
  5. 2026The Corroboration Illusion: When More News Makes LLM Forecasts Less Truecited by 0