Conic Formulations of Transport Metrics for Unbalanced Measure Networks and Hypernetworks

TL;DR

This paper introduces a new conic formulation of the Gromov-Wasserstein distance, extending it to compare general networks and hypernetworks, with theoretical properties and scalable algorithms demonstrated on complex datasets.

Contribution

It provides a novel semi-coupling formulation of the Conic Gromov-Wasserstein distance, extending its applicability and establishing fundamental properties and scalable algorithms.

Findings

The CGW metric is robust to measure perturbations.

The proposed algorithm converges and scales to high-dimensional data.

Experimental results validate the method on real-world datasets.

Abstract

The Gromov-Wasserstein (GW) variant of optimal transport, designed to compare probability densities defined over distinct metric spaces, has emerged as an important tool for the analysis of data with complex structure, such as ensembles of point clouds or networks. To overcome certain limitations, such as the restriction to comparisons of measures of equal mass and sensitivity to outliers, several unbalanced or partial transport relaxations of the GW distance have been introduced in the recent literature. This paper is concerned with the Conic Gromov-Wasserstein (CGW) distance introduced by S\'{e}journ\'{e}, Vialard, and Peyr\'{e}. We provide a novel formulation in terms of semi-couplings, and extend the framework beyond the metric measure space setting, to compare more general network and hypernetwork structures. With this new formulation, we establish several fundamental properties of the CGW metric, including its scaling behavior under dilation, variational convergence in the limit of volume growth constraints, and comparison bounds with established optimal transport metrics. We further derive quantitative bounds that characterize the robustness of the CGW metric to perturbations in the underlying measures. The hypernetwork formulation of CGW admits a simple and provably convergent block coordinate ascent algorithm for its estimation, and we demonstrate the computational tractability and scalability of our approach through experiments on synthetic and real-world high-dimensional and structured datasets.