GRANITE: A Generalized Regional Framework for Identifying Agreement in Feature-Based Explanations

Julia Herbinger, Gabriel Laberge, Maximilian Muschalik, Yann Pequignot, Marvin N. Wright, Fabian Fumagalli
Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, PMLR 300:4888-4896, 2026.

Abstract

Feature-based explanation methods aim to quantify how features influence the model’s behavior, either locally or globally, but different methods often disagree, producing conflicting explanations. This disagreement arises primarily from two sources: how feature interactions are handled and how feature dependencies are incorporated. We propose GRANITE, a generalized regional explanation framework that partitions the feature space into regions where interaction and distribution influences are minimized. This approach aligns different explanation methods, yielding more consistent and interpretable explanations. GRANITE unifies existing regional approaches, extends them to feature groups, and introduces a recursive partitioning algorithm to estimate such regions. We demonstrate its effectiveness on real-world datasets, providing a practical tool for consistent and interpretable feature explanations.

Cite this Paper

BibTeX
@InProceedings{pmlr-v300-herbinger26a, title = { GRANITE: A Generalized Regional Framework for Identifying Agreement in Feature-Based Explanations }, author = {Herbinger, Julia and Laberge, Gabriel and Muschalik, Maximilian and Pequignot, Yann and Wright, Marvin N. and Fumagalli, Fabian}, booktitle = {Proceedings of The 29th International Conference on Artificial Intelligence and Statistics}, pages = {4888--4896}, year = {2026}, editor = {Khan, Emtiyaz and Li, Yingzhen and Solin, Arno and Ramdas, Aaditya}, volume = {300}, series = {Proceedings of Machine Learning Research}, month = {02--05 May}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v300/main/assets/herbinger26a/herbinger26a.pdf}, url = {https://proceedings.mlr.press/v300/herbinger26a.html}, abstract = { Feature-based explanation methods aim to quantify how features influence the model’s behavior, either locally or globally, but different methods often disagree, producing conflicting explanations. This disagreement arises primarily from two sources: how feature interactions are handled and how feature dependencies are incorporated. We propose GRANITE, a generalized regional explanation framework that partitions the feature space into regions where interaction and distribution influences are minimized. This approach aligns different explanation methods, yielding more consistent and interpretable explanations. GRANITE unifies existing regional approaches, extends them to feature groups, and introduces a recursive partitioning algorithm to estimate such regions. We demonstrate its effectiveness on real-world datasets, providing a practical tool for consistent and interpretable feature explanations. } }
Endnote
%0 Conference Paper %T GRANITE: A Generalized Regional Framework for Identifying Agreement in Feature-Based Explanations %A Julia Herbinger %A Gabriel Laberge %A Maximilian Muschalik %A Yann Pequignot %A Marvin N. Wright %A Fabian Fumagalli %B Proceedings of The 29th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2026 %E Emtiyaz Khan %E Yingzhen Li %E Arno Solin %E Aaditya Ramdas %F pmlr-v300-herbinger26a %I PMLR %P 4888--4896 %U https://proceedings.mlr.press/v300/herbinger26a.html %V 300 %X Feature-based explanation methods aim to quantify how features influence the model’s behavior, either locally or globally, but different methods often disagree, producing conflicting explanations. This disagreement arises primarily from two sources: how feature interactions are handled and how feature dependencies are incorporated. We propose GRANITE, a generalized regional explanation framework that partitions the feature space into regions where interaction and distribution influences are minimized. This approach aligns different explanation methods, yielding more consistent and interpretable explanations. GRANITE unifies existing regional approaches, extends them to feature groups, and introduces a recursive partitioning algorithm to estimate such regions. We demonstrate its effectiveness on real-world datasets, providing a practical tool for consistent and interpretable feature explanations.
APA
Herbinger, J., Laberge, G., Muschalik, M., Pequignot, Y., Wright, M.N. & Fumagalli, F.. (2026). GRANITE: A Generalized Regional Framework for Identifying Agreement in Feature-Based Explanations . Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 300:4888-4896 Available from https://proceedings.mlr.press/v300/herbinger26a.html.

Related Material