Title:Interaction Tensor Shap
Abstract:Machine learning models have grown increasingly deep and high dimensional, making it difficult to understand how individual and combined features influence their predictions. While Shapley value based methods provide principled feature attributions, existing formulations cannot tractably evaluate higher order interactions: the Shapley Taylor Interaction Index (STII) requires exponential scale enumeration of subsets, and current tensor based approaches such as the Marginal SHAP Tensor (MST) are restricted to first order effects. The central problem is that no existing framework simultaneously preserves the axiomatic exactness of STII a…
Title:Interaction Tensor Shap
Abstract:Machine learning models have grown increasingly deep and high dimensional, making it difficult to understand how individual and combined features influence their predictions. While Shapley value based methods provide principled feature attributions, existing formulations cannot tractably evaluate higher order interactions: the Shapley Taylor Interaction Index (STII) requires exponential scale enumeration of subsets, and current tensor based approaches such as the Marginal SHAP Tensor (MST) are restricted to first order effects. The central problem is that no existing framework simultaneously preserves the axiomatic exactness of STII and avoids the exponential computational blow up inherent to high order discrete derivatives. Here we show that high order Shapley interactions can be represented exactly as tensor network contractions, enabling polynomial time and polylog depth computation under Tensor Train (TT) structure. We introduce Interaction Tensor SHAP (IT SHAP), which reformulates STII as the contraction of a Value Tensor and a Weight Tensor, and assume a finite state TT representation of the Weight Tensor with polynomial TT ranks. Under TT structured model and distribution tensors, we show that IT SHAP reduces the exponential complex Theta(4^n) of STII to NC2 parallel time. These results demonstrate that IT SHAP provides a unified, axiomatic, and computationally tractable formulation of main effects and higher order interactions in high dimensional models. This framework establishes a foundation for scalable interaction aware explainable AI, with implications for large black box models whose combinatorial structure has previously rendered interaction analysis infeasible.
| Comments: | 30 pages |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2512.05338 [cs.LG] |
| (or arXiv:2512.05338v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2512.05338 arXiv-issued DOI via DataCite |
Submission history
From: Hiroki Hasegawa [view email] [v1] Fri, 5 Dec 2025 00:34:15 UTC (48 KB)