2606.19410

Total: 1

#1 The Representational Limit of Scalar Interactions: An Interventional Decomposition [PDF] [Copy] [Kimi] [REL]

Authors: Potito Aghilar, Sabino Roccotelli, Stanislao Fidanza, Vito Walter Anelli, Sebastiano Stramaglia, Tommaso Di Noia

Signed pairwise interaction scores fundamentally conflate uniqueness (U), redundancy (R), and synergy (S). We prove this on a minimal 3-way XOR structural causal model: faithful indices such as Shapley-Taylor return zero per pair, whereas projective indices such as Shapley Interaction spread the third-order effect into pair scalars that conflate the three mechanisms. We introduce Stochastic Hi-Fi, a post-hoc, retraining-free predictability decomposition that estimates per-feature U/R/S profiles by interventional masked inference. The estimator provides exact interventional semantics, finite-sample Monte Carlo bounds, strict variance reduction from coupled diamond sampling, and uniform finite-vocabulary convergence. Across tabular SCMs, Stochastic Hi-Fi recovers structure missed by scalar baselines (up to 411x larger interaction-magnitude recovery ratios). It also separates redundant and synergistic heads in the GPT-2 IOI circuit. On NIH ChestX-ray14, Stochastic Hi-Fi matches GradCAM on Pointing Game and improves substantially on Deletion AUC.

Subjects: Machine Learning , Machine Learning

Publish: 2026-06-17 15:21:28 UTC