Implementation of Hierarchical Multi-Marginal Optimal Transport for Network Alignment in AAAI 2024.

prerequisites

• pot>=0.9.0
• numpy>=1.22.4
• scikit-learn>=1.1.2

code

• ./src/: source code to reproduce the experiments
  • hot_utils: utilities for hot, including calculating rwr, cross-graph cost, intra-graph cost, etc.
  • hot: the main HOT algorithm
  • log_mot: multi-marginal optimal transport in the log domain
  • run_ACM: Experiments on ACM and ACM(A) datasets
  • run_DBLP: Experiments on DBLP and DBLP(A) datasets
  • run_douban: experiments on Douban-230 dataset
• ./dataset/: dataset used for experiments

Simply run the run_ACM.py, run_DBLP.py, and run_douban.py to reproduce the experiment results.

If you find this paper helpful to your research, please kindly cite the following paper:

@misc{zeng2023hierarchical,
      title={Hierarchical Multi-Marginal Optimal Transport for Network Alignment}, 
      author={Zhichen Zeng and Boxin Du and Si Zhang and Yinglong Xia and Zhining Liu and Hanghang Tong},
      year={2023},
      eprint={2310.04470},
      archivePrefix={arXiv},
      primaryClass={cs.LG}
}