Research map / Learning theory and algorithms / Prediction and decision losses
On Locality of Local Explanation Models
Sahra Ghalebikesabi, Lucile Ter-Minassian, Karla DiazOrdaz, Christopher C. Holmes
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 small kernel bandwidth for reference sampling (Neighbourhood SHAP) instead of KernelSHAP | SHAP · adversarial OOD classifier test | AUC of OOD classifier | Improved |
| Use Neighbourhood SHAP instead of KernelSHAP | SHAP · COMPAS adversarial black box | relative attribution to race | Improved |
| Use smaller bandwidth in Neighbourhood SHAP | Neighbourhood SHAP · four datasets with XGBoost | MSE of local prediction | Improved |
| Use small instead of large kernel width | Neighbourhood SHAP · bike data | temperature attribution sign | Mixed |
| Use small instead of large kernel width | Neighbourhood SHAP · Boston data | LSTAT attribution sign: correlation -0.76 vs +0.15 | Mixed |
| Apply Neighbourhood SHAP to DeepSHAP instead of standard DeepSHAP | DeepSHAP · MNIST digit 8 | attribution sparsity | Improved |
| Apply Neighbourhood DeepSHAP instead of DeepSHAP | DeepSHAP · MNIST digit 8 | change in log odds of predicting 3 | No clear effect |
| Apply smoothing (Smoothed DeepSHAP) instead of DeepSHAP | DeepSHAP · MNIST digit 8 | change in log odds | Hurt |
| Fewer features in data set (e.g. iris) for bandwidth comparison | Neighbourhood SHAP · four datasets | MSE gap between bandwidths | Mixed |
| Increase bandwidth in Smoothed SHAP | Smoothed SHAP · Adult data | attribution of age and education | Mixed |
Closest papers by meaning
- Regional Explanations: Bridging Local and Global Variable Importance (NeurIPS 2025 · similarity 0.89)
- SHAP values via sparse Fourier representation (NeurIPS 2025 · similarity 0.86)
- Helpful or Harmful Data? Fine-tuning-free Shapley Attribution for Explaining Language Model Predictions (ICML 2024 · similarity 0.86)
- shapiq: Shapley Interactions for Machine Learning (NeurIPS 2024 · similarity 0.86)
- A Unified Framework for Provably Efficient Algorithms to Estimate Shapley Values (NeurIPS 2025 · similarity 0.86)
- InstaSHAP: Interpretable Additive Models Explain Shapley Values Instantly (ICLR 2025 · similarity 0.86)
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.
