English

Pixel-level Certified Explanations via Randomized Smoothing

Machine Learning 2025-06-19 v1 Artificial Intelligence Computer Vision and Pattern Recognition

Abstract

Post-hoc attribution methods aim to explain deep learning predictions by highlighting influential input pixels. However, these explanations are highly non-robust: small, imperceptible input perturbations can drastically alter the attribution map while maintaining the same prediction. This vulnerability undermines their trustworthiness and calls for rigorous robustness guarantees of pixel-level attribution scores. We introduce the first certification framework that guarantees pixel-level robustness for any black-box attribution method using randomized smoothing. By sparsifying and smoothing attribution maps, we reformulate the task as a segmentation problem and certify each pixel's importance against 2\ell_2-bounded perturbations. We further propose three evaluation metrics to assess certified robustness, localization, and faithfulness. An extensive evaluation of 12 attribution methods across 5 ImageNet models shows that our certified attributions are robust, interpretable, and faithful, enabling reliable use in downstream tasks. Our code is at https://github.com/AlaaAnani/certified-attributions.

Keywords

Cite

@article{arxiv.2506.15499,
  title  = {Pixel-level Certified Explanations via Randomized Smoothing},
  author = {Alaa Anani and Tobias Lorenz and Mario Fritz and Bernt Schiele},
  journal= {arXiv preprint arXiv:2506.15499},
  year   = {2025}
}
R2 v1 2026-07-01T03:23:41.392Z