AKE

The implementation of our ICDM 2019 paper "Guiding Entity Alignment via Adversarial Knowledge Embedding" AKE.

Requirements

python 3.5.3
torch == 0.3.1
numpy == 1.15
scipy == 1.1.0
scikit-learn == 0.20.0

How to use

Dataset

tar -zxvf data.tar.gz data
The data folder includes our propocessed data JA-EN for training and testing.
The orginal datasets can be founded from here.

Training

zsh train.sh/train_variants.sh # training AKE and variants with default hyper-parameters

Testing

python test.py

Citation

If you find the code is useful for your research, please cite this paper:

@inproceedings{AKE
author={Xixun Lin and Hong Yang and Jia Wu and Chuan Zhou and Bin Wang},
title={Guiding Entity Alignment via Adversarial Knowledge Embedding},
journal={IEEE International Conference On Data Mining (ICDM)},
year={2019}
}