Structure-Preserving Graph Representation Learning（SPGRL）

The official code of ICDM2022 paper: [Structure-Preserving Graph Representation Learning]
Ruiyi Fang, Liangjian Wen, Zhao Kang, Jianzhuang Liu

We propose a new Structure-Preserving Graph Representation Learning method called SPGRL. Our main idea is maximizing the mutual information (MI) between the graph structure and feature embedding.

The module is illustrated as follows:

Environment Settings

• python == 3.8
• Pytorch == 1.1.0
• Numpy == 1.16.2
• SciPy == 1.3.1
• Networkx == 2.4
• scikit-learn == 0.21.3

Usage

python main.py -d dataset -l labelrate

• dataset: including [citeseer, uai, acm, BlogCatalog, flickr, cora], required.
• labelrate: including [20, 40, 60], required.

e.g.

python main.py -d citeseer -l 20