Sum-of-Parts: Self-Attributing Neural Networks with End-to-End Learning of Feature Groups

Weiqiu You, Helen Qu, Marco Gatti, Bhuvnesh Jain, Eric Wong
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:72747-72785, 2025.

Abstract

Self-attributing neural networks (SANNs) present a potential path towards interpretable models for high-dimensional problems, but often face significant trade-offs in performance. In this work, we formally prove a lower bound on errors of per-feature SANNs, whereas group-based SANNs can achieve zero error and thus high performance. Motivated by these insights, we propose Sum-of-Parts (SOP), a framework that transforms any differentiable model into a group-based SANN, where feature groups are learned end-to-end without group supervision. SOP achieves state-of-the-art performance for SANNs on vision and language tasks, and we validate that the groups are interpretable on a range of quantitative and semantic metrics. We further validate the utility of SOP explanations in model debugging and cosmological scientific discovery.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-you25e, title = {Sum-of-Parts: Self-Attributing Neural Networks with End-to-End Learning of Feature Groups}, author = {You, Weiqiu and Qu, Helen and Gatti, Marco and Jain, Bhuvnesh and Wong, Eric}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {72747--72785}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/you25e/you25e.pdf}, url = {https://proceedings.mlr.press/v267/you25e.html}, abstract = {Self-attributing neural networks (SANNs) present a potential path towards interpretable models for high-dimensional problems, but often face significant trade-offs in performance. In this work, we formally prove a lower bound on errors of per-feature SANNs, whereas group-based SANNs can achieve zero error and thus high performance. Motivated by these insights, we propose Sum-of-Parts (SOP), a framework that transforms any differentiable model into a group-based SANN, where feature groups are learned end-to-end without group supervision. SOP achieves state-of-the-art performance for SANNs on vision and language tasks, and we validate that the groups are interpretable on a range of quantitative and semantic metrics. We further validate the utility of SOP explanations in model debugging and cosmological scientific discovery.} }
Endnote
%0 Conference Paper %T Sum-of-Parts: Self-Attributing Neural Networks with End-to-End Learning of Feature Groups %A Weiqiu You %A Helen Qu %A Marco Gatti %A Bhuvnesh Jain %A Eric Wong %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-you25e %I PMLR %P 72747--72785 %U https://proceedings.mlr.press/v267/you25e.html %V 267 %X Self-attributing neural networks (SANNs) present a potential path towards interpretable models for high-dimensional problems, but often face significant trade-offs in performance. In this work, we formally prove a lower bound on errors of per-feature SANNs, whereas group-based SANNs can achieve zero error and thus high performance. Motivated by these insights, we propose Sum-of-Parts (SOP), a framework that transforms any differentiable model into a group-based SANN, where feature groups are learned end-to-end without group supervision. SOP achieves state-of-the-art performance for SANNs on vision and language tasks, and we validate that the groups are interpretable on a range of quantitative and semantic metrics. We further validate the utility of SOP explanations in model debugging and cosmological scientific discovery.
APA
You, W., Qu, H., Gatti, M., Jain, B. & Wong, E.. (2025). Sum-of-Parts: Self-Attributing Neural Networks with End-to-End Learning of Feature Groups. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:72747-72785 Available from https://proceedings.mlr.press/v267/you25e.html.

Related Material