Research map / Label-efficient and robust learning / Recognition and long-tail
Explaining Image Classifiers using Statistical Fault Localization
Youcheng Sun, Hana Chockler, Xiaowei Huang, Daniel Kroening
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use DeepCover (SFL ranking) instead of Extremal for masking until class change | DeepCover vs Extremal · ImageNet VGG16 | masked pixels to change class: 4% -> 2% | Improved |
| use DeepCover instead of Extremal on the Chimera benchmark | DeepCover vs Extremal · Chimera red panda | detection rate (IoU >= 0.5): 76.7% | Improved |
| use DeepCover instead of Extremal for explanation size | DeepCover vs Extremal · ImageNet VGG16 | explanation size: 40% of inputs within 10% of pixels | Improved |
| reduce the mutant test suite from 2,000 to 200 (ablation) | DeepCover · ImageNet | explanation size: +3% | Hurt |
| use DeepCover instead of RISE | DeepCover vs RISE · ImageNet | explanation size | Improved |
| use DeepCover instead of GradCAM on Chimera | DeepCover vs GradCAM · Chimera red panda | detection rate | Improved |
Closest papers by meaning
- Explanations for Occluded Images (ICCV 2021 · similarity 0.86)
- Explanation-based Data Augmentation for Image Classification (NeurIPS 2021 · similarity 0.85)
- Global Explanations for Image Classifiers (Student Abstract) (AAAI 2023 · similarity 0.84)
- Understanding deep learning requires rethinking generalization (ICLR 2017 · similarity 0.84)
- Deep Probabilistic Supervision for Image Classification (NeurIPS 2026 · similarity 0.84)
- Q-SENN: Quantized Self-Explaining Neural Networks (AAAI 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
