www2023tiger

Codes for WWW 2023 paper "TIGER: Temporal Interaction Graph Embedding with Restarts" arXiv

Data

Please download data from the project homepage of JODIE and pre-process them with the script provided by TGN.

How to use

Temporal Link Prediction

python train_self_supervised.py --data [DATA] --msg_src [left/right] --upd_src [left/right] --restarter [seq/static] --restart_prob [0/0.001/...]

If you want to use mooc/lastfm datasets, please pass one more argument: --dim 100.

Temporal Link Prediction with multi-GPU

python train_self_supervised_ddp.py --gpu 0,1,2,3 [...and other arguments]

Node Classification

python train_supervised.py --code [CODE]

Here, [CODE] is the HASH code of a trained model with train_self_supervised.py.

Cite

@inproceedings{zhang2023tiger,
  title={TIGER: Temporal Interaction Graph Embedding with Restarts},
  author={Zhang, Yao and Xiong, Yun and Liao, Yongxiang and Sun, Yiheng and Jin, Yucheng and Zheng, Xuehao and Zhu, Yangyong},
  booktitle={ACM Web Conference},
  year={2023}
}