Research map / Label-efficient and robust learning / Recognition and long-tail
VeriX: Towards Verified Explainability of Deep Neural Networks
Min Wu, Haoze Wu, Clark Barrett
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 |
|---|---|---|---|
| sensitivity traversal instead of random traversal (ablation) | VeriX · MNIST | explanation size: 294.2 -> 180.6 | Improved |
| sensitivity traversal instead of random traversal (ablation) | VeriX · GTSRB | explanation size: 383.5 -> 357.0 | Improved |
| sensitivity traversal instead of random traversal (ablation) | VeriX · MNIST | generation time: 157.47 -> 174.77 | Hurt |
| sensitivity traversal instead of random traversal (ablation) | VeriX · GTSRB | generation time: 814.18 -> 853.91 | Hurt |
| VeriX instead of Anchors | explanation method · MNIST | explanation size: 494.9 -> 180.6 | Improved |
| VeriX instead of LIME | explanation method · MNIST | explanation size: 432.8 -> 180.6 | Improved |
| VeriX instead of Anchors | explanation method · GTSRB | explanation size: 557.7 -> 357.0 | Improved |
| VeriX instead of Anchors | explanation method · MNIST | generation time: 13.46 -> 174.77 | Hurt |
| VeriX instead of Anchors | explanation method · GTSRB | robustness to 1000 perturbations: 15% -> 100% | Improved |
| VeriX instead of Anchors | explanation method · MNIST | robustness to 1000 perturbations: 37% -> 100% | Improved |
| incomplete (DeepPoly) instead of complete verification (ablation) | VeriX sensitivity · MNIST | explanation size: 180.6 -> 330.8 | Hurt |
| incomplete (DeepPoly) instead of complete verification (ablation) | VeriX sensitivity · MNIST | generation time: 174.77 -> 106.11 | Improved |
Closest papers by meaning
- CF-OPT: Counterfactual Explanations for Structured Prediction (ICML 2024 · similarity 0.86)
- Towards Modeling Uncertainties of Self-Explaining Neural Networks via Conformal Prediction (AAAI 2024 · similarity 0.85)
- Explaining, Fast and Slow: Abstraction and Refinement of Provable Explanations (ICML 2025 · similarity 0.85)
- Rethinking Visual Counterfactual Explanations Through Region Constraint (ICLR 2025 · similarity 0.85)
- Explainable Neural Networks with Guarantee: A Sparse Estimation Approach (AAAI 2025 · similarity 0.85)
- Q-SENN: Quantized Self-Explaining Neural Networks (AAAI 2024 · similarity 0.85)
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.
