Research map / Label-efficient and robust learning / Recognition and long-tail
Understanding Global Feature Contributions With Additive Importance Measures
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 SAGE instead of the strongest baseline, Permutation Test | Global feature importance · Bank Marketing | correlation between total importance and performance: 0.9918 -> 0.9937 | Helped |
| Use SAGE instead of the strongest baseline, Permutation Test | Global feature importance · Bike Demand | correlation between total importance and performance: 0.9798 -> 0.9815 | Helped |
| Use SAGE instead of the strongest baseline, Permutation Test | Global feature importance · Credit Quality | correlation between total importance and performance: 0.9571 -> 0.9565 | Hurt |
| Use SAGE instead of the strongest baseline, Univariate | Global feature importance · BRCA | correlation between total importance and performance: 0.8718 -> 0.8611 | Hurt |
| Use SAGE instead of the strongest baseline, Univariate | Global feature importance · MNIST | correlation between total importance and performance: 0.4865 -> 0.4868 | Helped |
| Use sampling-based SAGE rather than averaging local SHAP explanations | Global explanation computation · Experiments across datasets | model evaluations required for the same values: 2-4 orders of magnitude more model evaluations | Helped |
Closest papers by meaning
- Testing Semantic Importance via Betting (NeurIPS 2024 · similarity 0.85)
- Feature Importance Metrics in the Presence of Missing Data (ICML 2025 · similarity 0.84)
- Accurate Estimation of Feature Importance Faithfulness for Tree Models (AAAI 2025 · similarity 0.84)
- Local vs. Global Interpretability: A Computational Complexity Perspective (ICML 2024 · similarity 0.83)
- Regional Explanations: Bridging Local and Global Variable Importance (NeurIPS 2025 · similarity 0.83)
- Conditional Feature Importance with Generative Modeling Using Adversarial Random Forests (AAAI 2025 · similarity 0.83)
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.