Skip to content
master
Go to file
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.

README.md

S-GWL

This package includes the implementation of my NeurIPS2019 work "Scalable Gromov-Wasserstein Learning for Graph Partitioning and Matching" [https://arxiv.org/pdf/1905.07645.pdf]

The examples include:

  • Partition a single graph (i.e., community detection)
  • Match two or more graphs (i.e., network alignment)

The package is developed on Ubuntu 18.04. Specifically, the baseline "metis" has been compiled as a library "libmetis.so" in the "baselines/metis-5.1.0" folder. The source code can be found at http://glaros.dtc.umn.edu/gkhome/metis/metis/download.

Dependencies

Note that after install the wrapper "metis", you need to set the path of libmetis.so manually in metis.py.

Baselines

Graph partitioning

  • Metis
  • Louvain
  • FastGreedy
  • Fluid

Graph matching

We download these two packages from https://ttic.uchicago.edu/~hashemifar/. Please read the "readme" file in their folders to run these two methods.

Citations

@article{xu2019scalable, title={Scalable Gromov-Wasserstein Learning for Graph Partitioning and Matching}, author={Xu, Hongteng and Luo, Dixin and Carin, Lawrence}, journal={arXiv preprint arXiv:1905.07645}, year={2019} }

About

Scalable Gromov-Wasserstein Learning for Graph Partitioning and Matching

Resources

Releases

No releases published

Packages

No packages published
You can’t perform that action at this time.