remorax / emgcn

This is an implementation of the paper Entity Alignment for Knowledge Graphs with Multi-order Convolutional Networks

EMGCN

Code of the paper: Entity Alignment for Knowledge Graphs with Multi-order Convolutional Networks.

Environment

• python>=3.5
• networkx == 1.11 (important!)
• pytorch >= 1.2.0
• numpy >= 1.18.1

Dataset

You can download our processed dataset from: https://drive.google.com/file/d/12XL08tB8zplCNhzLE-9qbsFFum7RoV6r/view?usp=sharing.

Running

python -u network_alignment.py --dataset_name zh_en --source_dataset data/networkx/zh_enDI/zh/graphsage/ --target_dataset data/networkx/zh_enDI/en/graphsage --groundtruth data/networkx/zh_enDI/dictionaries/groundtruth EMGCN --sparse --log 

Citation

Please politely cite our work as follows:

Nguyen Thanh Tam, Huynh Thanh Trung, Hongzhi Yin, Tong Van Vinh, Darnbi Sakong, Bolong Zheng, Nguyen Quoc Viet Hung. Entity Alignment for Knowledge Graphs with Multi-order Convolutional Networks . In: TKDE 2020.