Research map / Learning theory and algorithms / Prediction and decision losses
A Unifying Framework to the Analysis of Interaction Methods using Synergy Functions
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 |
|---|---|---|---|
| Replace Shapley-Taylor-style distribution with Integrated Hessian (augmented) distribution of monomial synergies | gradient-based interaction methods on analytic functions · theoretical analysis (no experiments) | axiom satisfaction (baseline test for interactions) | Mixed |
| Use the augmented Integrated Hessian (pre-distributing single and pairwise synergies) instead of plain Integrated Hessian | Integrated Hessian · theoretical analysis (no experiments) | baseline test for interactions (k ≤ n) satisfied | Helped |
| Use the Sum of Powers top-distributing gradient-based scheme instead of recursive Integrated Hessian distribution | gradient-based kth-order interaction methods · theoretical analysis (no experiments) | satisfies baseline test for interactions and top-distributing property | Helped |
| Use Recursive Shapley (not top-distributing) instead of Shapley-Taylor for kth-order interactions | binary feature kth-order interactions · theoretical analysis (no experiments) | baseline test for interactions (k ≤ n) | Hurt |
| Use Integrated Gradients (gradient-based) instead of Shapley value to distribute synergies of monomials | feature attribution for monomial synergies · worked example x1^100 x2 synergy | attribution fairness by monomial form | Helped |
Closest papers by meaning
- Explaining Probabilistic Models with Distributional Values (ICML 2024 · similarity 0.87)
- NfgTransformer: Equivariant Representation Learning for Normal-form Games (ICLR 2024 · similarity 0.85)
- Unlocking the Game: Estimating Games in Möbius Representation for Explanation and High-Order Interaction Detection (AAAI 2025 · similarity 0.85)
- Towards Attributions of Input Variables in a Coalition (ICML 2025 · similarity 0.85)
- Understanding Model Selection for Learning in Strategic Environments (NeurIPS 2024 · similarity 0.84)
- Distributional Preference Learning: Understanding and Accounting for Hidden Context in RLHF (ICLR 2024 · similarity 0.84)
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.