Research map / Learning theory and algorithms / Prediction and decision losses
Consistent Sufficient Explanations and Minimal Local Rules for explaining the decision of any classifier or regressor
Salim I. Amoukou, Nicolas J-B. Brunel
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 Sufficient Explanation instead of SHAP for feature discovery | Sufficient Explanation · Synthetic switch regression | TPR: 73% -> 99% | Improved |
| Use Sufficient Explanation instead of SHAP for FDR | Sufficient Explanation · Synthetic switch regression | FDR: 27% -> 2% | Improved |
| Use Sufficient Explanation instead of LIME | Sufficient Explanation · Synthetic switch regression | TPR: 50% -> 99% | Improved |
| Use Sufficient Explanation instead of INVASE | Sufficient Explanation · Synthetic switch regression | FDR: 87% -> 2% | Improved |
| Use Sufficient Explanation instead of INVASE for P-MSE | Sufficient Explanation · Synthetic switch regression | P-MSE: 0.006 -> 0.02 | Hurt |
| Use Sufficient Explanation instead of SHAP for P-MSE | Sufficient Explanation · Synthetic switch regression | P-MSE: 0.79 -> 0.02 | Improved |
| Use LXI instead of SHAP attribution | LXI · Synthetic model, X5>0 | attribution of X1: 1.47 -> 0 | Improved |
| Replace Anchors with Sufficient Rules | Sufficient Rules · COMPAS | Coverage: 0.23 -> 0.30 | Improved |
| Replace Anchors with Sufficient Rules | Sufficient Rules · NHANESI | Coverage: 0.31 -> 0.41 | Improved |
| Replace Anchors with Sufficient Rules | Sufficient Rules · ATTRITION | Correctness: 0.95 -> 0.95 | No clear effect |
| Replace Anchors with Sufficient Rules | Sufficient Rules · COMPAS | Correctness: 0.92 -> 0.95 | Improved |
| Replace Anchors with Sufficient Rules for rule length | Sufficient Rules · NHANESI | max rule size: 16 -> 7 | Improved |
Closest papers by meaning
- Computing Probabilistic Explanations for ML Models: Fixed-Parameter Algorithms (AAAI 2026 · similarity 0.88)
- Unifying Formal Explanations: A Complexity-Theoretic Perspective (ICLR 2026 · similarity 0.88)
- Probabilistic Explanations for Linear Models (AAAI 2025 · similarity 0.87)
- Explainable Neural Networks with Guarantee: A Sparse Estimation Approach (AAAI 2025 · similarity 0.86)
- Explain Yourself, Briefly! Self-Explaining Neural Networks with Concise Sufficient Reasons (ICLR 2025 · similarity 0.85)
- Improving Decision Sparsity (NeurIPS 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.
