Homological Representation Learning for Molecular Graphs

Yoshihiro Maruyama, Arisa Yasuda
Proceedings of the 4th (2025) and 3rd (2024) NeurIPS Workshops on Symmetry and Geometry in Neural Representations, PMLR 282:369-380, 2026.

Abstract

We propose Homological Representation Learning (HomRL), an architecture-agnostic regularization method for graph encoders that aligns latent embeddings with an efficiently computable homological signature of the input. In this paper, we give both theoretical results on representation invariance bounds and empirical results on molecular graph classification tasks.

Cite this Paper

BibTeX
@InProceedings{pmlr-v282-maruyama26a, title = {Homological Representation Learning for Molecular Graphs}, author = {Maruyama, Yoshihiro and Yasuda, Arisa}, booktitle = {Proceedings of the 4th (2025) and 3rd (2024) NeurIPS Workshops on Symmetry and Geometry in Neural Representations}, pages = {369--380}, year = {2026}, editor = {Acosta, Francisco and Azeglio, Simone and Tolooshams, Bahareh and van de Geijn, Chase and Shewmake, Christian and Sanborn, Sophia and Miolane, Nina}, volume = {282}, series = {Proceedings of Machine Learning Research}, month = {14 Dec 2024--07 Dec 2025}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v282/main/assets/maruyama26a/maruyama26a.pdf}, url = {https://proceedings.mlr.press/v282/maruyama26a.html}, abstract = {We propose Homological Representation Learning (HomRL), an architecture-agnostic regularization method for graph encoders that aligns latent embeddings with an efficiently computable homological signature of the input. In this paper, we give both theoretical results on representation invariance bounds and empirical results on molecular graph classification tasks.} }
Endnote
%0 Conference Paper %T Homological Representation Learning for Molecular Graphs %A Yoshihiro Maruyama %A Arisa Yasuda %B Proceedings of the 4th (2025) and 3rd (2024) NeurIPS Workshops on Symmetry and Geometry in Neural Representations %C Proceedings of Machine Learning Research %D 2026 %E Francisco Acosta %E Simone Azeglio %E Bahareh Tolooshams %E Chase van de Geijn %E Christian Shewmake %E Sophia Sanborn %E Nina Miolane %F pmlr-v282-maruyama26a %I PMLR %P 369--380 %U https://proceedings.mlr.press/v282/maruyama26a.html %V 282 %X We propose Homological Representation Learning (HomRL), an architecture-agnostic regularization method for graph encoders that aligns latent embeddings with an efficiently computable homological signature of the input. In this paper, we give both theoretical results on representation invariance bounds and empirical results on molecular graph classification tasks.
APA
Maruyama, Y. & Yasuda, A.. (2026). Homological Representation Learning for Molecular Graphs. Proceedings of the 4th (2025) and 3rd (2024) NeurIPS Workshops on Symmetry and Geometry in Neural Representations, in Proceedings of Machine Learning Research 282:369-380 Available from https://proceedings.mlr.press/v282/maruyama26a.html.

Related Material