This is the project for the following paper, accepted in proceeding for 30th The Web Conference 2021, Ljubljana, Slovenia. You can cite for now.
@article{zhang2020tg,
title={TG-GAN: Deep Generative Models for Continuously-time Temporal Graph Generation},
author={Zhang, Liming and Zhao, Liang and Qin, Shan and Pfoser, Dieter},
journal={arXiv preprint arXiv:2005.08323},
year={2020}
}
The main training and inference codes for different datasets are in main_*.py scripts.
There is also codes developed for dynamic graph metric in MMD distance evaluation.
The continuous-time graph metrics are in evaluation.py use another library tacoma,
and the folder continuous_time_evaluation_and_DSBM_matlab contains the discrete-time
graph metrics and also DSBM models. The referred libraries can be found in the folder too.
Please cite this paper properly if you need to use the evaluation codes.