Scalable Gromov-Wasserstein Learning for Graph Partitioning and Matching

Abstract

We propose a scalable Gromov-Wasserstein learning (S-GWL) method and establish a novel and theoretically-supported paradigm for large-scale graph analysis. The proposed method is based on the fact that Gromov-Wasserstein discrepancy is a pseudometric on graphs. Given two graphs, the optimal transport associated with their Gromov-Wasserstein discrepancy provides the correspondence between their nodes and achieves graph matching. When one of the graphs is a predefined graph with isolated but self-connected nodes ($i.e.$, disconnected graph), the optimal transport indicates the clustering structure of the other graph and achieves graph partitioning. Further, we extend our method to multi-graph partitioning and matching by learning a Gromov-Wasserstein barycenter graph for multiple observed graphs. Our method combines a recursive $K$-partition mechanism with a warm-start proximal gradient algorithm, whose time complexity is $\mathcal{O}(K(E+V)\log_K V)$ for graphs with $V$ nodes and $E$ edges. To our knowledge, our method is the first attempt to make Gromov-Wasserstein discrepancy applicable to large-scale graph analysis and unify graph partitioning and matching into the same framework. It outperforms state-of-the-art graph partitioning and matching methods, achieving a trade-off between accuracy and efficiency.

Cite

Text

Xu et al. "Scalable Gromov-Wasserstein Learning for Graph Partitioning and Matching." Neural Information Processing Systems, 2019.

Markdown

[Xu et al. "Scalable Gromov-Wasserstein Learning for Graph Partitioning and Matching." Neural Information Processing Systems, 2019.](https://mlanthology.org/neurips/2019/xu2019neurips-scalable/)

BibTeX

@inproceedings{xu2019neurips-scalable,
  title     = {{Scalable Gromov-Wasserstein Learning for Graph Partitioning and Matching}},
  author    = {Xu, Hongteng and Luo, Dixin and Carin, Lawrence},
  booktitle = {Neural Information Processing Systems},
  year      = {2019},
  pages     = {3052-3062},
  url       = {https://mlanthology.org/neurips/2019/xu2019neurips-scalable/}
}