Scalable Sobolev IPM for Probability Measures on a Graph

Tam Le, Truyen Nguyen, Hideitsu Hino, Kenji Fukumizu
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:32771-32802, 2025.

Abstract

We investigate the Sobolev IPM problem for probability measures supported on a graph metric space. Sobolev IPM is an important instance of integral probability metrics (IPM), and is obtained by constraining a critic function within a unit ball defined by the Sobolev norm. In particular, it has been used to compare probability measures and is crucial for several theoretical works in machine learning. However, to our knowledge, there are no efficient algorithmic approaches to compute Sobolev IPM effectively, which hinders its practical applications. In this work, we establish a relation between Sobolev norm and weighted $L^p$-norm, and leverage it to propose a novel regularization for Sobolev IPM. By exploiting the graph structure, we demonstrate that the regularized Sobolev IPM provides a closed-form expression for fast computation. This advancement addresses long-standing computational challenges, and paves the way to apply Sobolev IPM for practical applications, even in large-scale settings. Additionally, the regularized Sobolev IPM is negative definite. Utilizing this property, we design positive-definite kernels upon the regularized Sobolev IPM, and provide preliminary evidences of their advantages for comparing probability measures on a given graph for document classification and topological data analysis.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-le25a, title = {Scalable Sobolev {IPM} for Probability Measures on a Graph}, author = {Le, Tam and Nguyen, Truyen and Hino, Hideitsu and Fukumizu, Kenji}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {32771--32802}, 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/le25a/le25a.pdf}, url = {https://proceedings.mlr.press/v267/le25a.html}, abstract = {We investigate the Sobolev IPM problem for probability measures supported on a graph metric space. Sobolev IPM is an important instance of integral probability metrics (IPM), and is obtained by constraining a critic function within a unit ball defined by the Sobolev norm. In particular, it has been used to compare probability measures and is crucial for several theoretical works in machine learning. However, to our knowledge, there are no efficient algorithmic approaches to compute Sobolev IPM effectively, which hinders its practical applications. In this work, we establish a relation between Sobolev norm and weighted $L^p$-norm, and leverage it to propose a novel regularization for Sobolev IPM. By exploiting the graph structure, we demonstrate that the regularized Sobolev IPM provides a closed-form expression for fast computation. This advancement addresses long-standing computational challenges, and paves the way to apply Sobolev IPM for practical applications, even in large-scale settings. Additionally, the regularized Sobolev IPM is negative definite. Utilizing this property, we design positive-definite kernels upon the regularized Sobolev IPM, and provide preliminary evidences of their advantages for comparing probability measures on a given graph for document classification and topological data analysis.} }
Endnote
%0 Conference Paper %T Scalable Sobolev IPM for Probability Measures on a Graph %A Tam Le %A Truyen Nguyen %A Hideitsu Hino %A Kenji Fukumizu %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-le25a %I PMLR %P 32771--32802 %U https://proceedings.mlr.press/v267/le25a.html %V 267 %X We investigate the Sobolev IPM problem for probability measures supported on a graph metric space. Sobolev IPM is an important instance of integral probability metrics (IPM), and is obtained by constraining a critic function within a unit ball defined by the Sobolev norm. In particular, it has been used to compare probability measures and is crucial for several theoretical works in machine learning. However, to our knowledge, there are no efficient algorithmic approaches to compute Sobolev IPM effectively, which hinders its practical applications. In this work, we establish a relation between Sobolev norm and weighted $L^p$-norm, and leverage it to propose a novel regularization for Sobolev IPM. By exploiting the graph structure, we demonstrate that the regularized Sobolev IPM provides a closed-form expression for fast computation. This advancement addresses long-standing computational challenges, and paves the way to apply Sobolev IPM for practical applications, even in large-scale settings. Additionally, the regularized Sobolev IPM is negative definite. Utilizing this property, we design positive-definite kernels upon the regularized Sobolev IPM, and provide preliminary evidences of their advantages for comparing probability measures on a given graph for document classification and topological data analysis.
APA
Le, T., Nguyen, T., Hino, H. & Fukumizu, K.. (2025). Scalable Sobolev IPM for Probability Measures on a Graph. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:32771-32802 Available from https://proceedings.mlr.press/v267/le25a.html.

Related Material