Research map / Learning theory and algorithms / Prediction and decision losses
Faith-Shap: The Faithful Shapley Interaction Index
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 |
|---|---|---|---|
| Estimate interaction attributions with weighted least squares rather than permutation-based sampling | Faith-Shap · Simplified IMDB and Bank marketing | model evaluations to reach averaged squared distance < 10^-3: 2781.2 vs 887.4; 7368.3 vs 893.7 | Helped |
| Use faithfulness-constrained interaction attributions instead of Shapley Taylor interaction distribution | Faith-Shap · Diminishing marginal utility game (p=0.1, ℓ=2) | order-1 and order-2 interaction scores: 0/0.1 -> 0.95/-0.091 | Helped |
| Use faithfulness-constrained interaction attributions instead of Shapley Taylor interaction distribution | Faith-Shap · Diminishing marginal utility game (p=0.2, ℓ=2) | order-1 and order-2 interaction scores: 0/0 -> 0.95/-0.191 | Helped |
| Use faithful interaction attributions rather than Shapley Taylor | Faith-Shap · Increasing marginal utility game (ℓ=2) | order-2 interaction score: -0.29 -> 0.07 | Helped |
| Make interactions depend on the maximum order and fit coalition values by weighted regression | Faith-Shap · Simplified IMDB and synthetic sparse functions | estimation accuracy | Helped |
Closest papers by meaning
- shapiq: Shapley Interactions for Machine Learning (NeurIPS 2024 · similarity 0.80)
- KernelSHAP-IQ: Weighted Least Square Optimization for Shapley Interactions (ICML 2024 · similarity 0.80)
- Proxy-Based Approximation of Shapley and Banzhaf Interactions (NeurIPS 2026 · similarity 0.78)
- Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles (AAAI 2024 · similarity 0.77)
- AppWorld-UL: Benchmarking Diverse Agent-User Interactions for Tool-Use (ICML 2026 · similarity 0.77)
- Measuring and Mitigating the Distributional Gap Between Real and Simulated User Behaviors (NeurIPS 2026 · similarity 0.77)
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.