A reference implementation for "Collective Multi-type Entity Alignment Between Knowledge Graphs", published in WWW 2020.

If you find our paper useful, please consider cite the following paper

author = {Zhu, Qi and Wei, Hao and Sisman, Bunyamin and Zheng, Da and Faloutsos, Christos and Dong, Xin Luna and Han, Jiawei},
title = {Collective Multi-Type Entity Alignment Between Knowledge Graphs},
year = {2020},
url = {},
doi = {10.1145/3366423.3380289},
booktitle = {Proceedings of The Web Conference 2020}


Unfortunately, the original data used is not public available. But this reference implementation could be easily adopt to structured data: knowledge graph, knowledge base and etc. See examples below for details.

We are collecting more public available knowledge graphs, stay tuned! Feel free to contact me ([email protected]) if you want to add your dataset in this repository.


pip install -r requirements.txt

Run the code

Prepare the pre-trained fastText embedding

Most of the attributes in a knowledge graph is text. Obtain your binarized pre-trained word embeddings $PATH at fastText. I'm using enwiki9.bin

python --gpu=0 --pretrain-path=$PATH