The code for our ICLR 2024 paper: "Beyond Spatio-Temporal Representations: Evolving Fourier Transform for Temporal Graphs" (https://openreview.net/forum?id=uvFhCUPjtI). We have implemented our methods in Pytorch.
- Python 3.7
- torch 1.11.0
- dgl 0.8.2
You need to run the file new_data.py to generate the data format needed for the model. The detailed commands
can be found in load_{dataset}.sh
You need to run the file generate_neg.py to generate data to speed up the test. You can set the data set in the file.
Then you can run the file new_main.py to train and test our model on the SR datasets.
The detailed commands can be found in load_{dataset}.sh
If our research has benefited you, please cite:
@inproceedings{
bastos2024beyond,
title={Beyond Spatio-Temporal Representations: Evolving Fourier Transform for Temporal Graphs},
author={Anson Bastos and Kuldeep Singh and Abhishek Nadgeri and Manish Singh and Toyotaro Suzumura},
booktitle={The Twelfth International Conference on Learning Representations},
year={2024},
url={https://openreview.net/forum?id=uvFhCUPjtI}
}