SURGE: Sequential Recommendation with Graph Neural Networks

This is our TensorFlow implementation for the paper:

Sequential Recommendation with Graph Neural Networks. SIGIR '21: Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval.

Please cite our paper if you use this repository.

@inproceedings{chang2021sequential,
  title={Sequential Recommendation with Graph Neural Networks},
  author={Chang, Jianxin and Gao, Chen and Zheng, Yu and Hui, Yiqun and Niu, Yanan and Song, Yang and Jin, Depeng and Li, Yong},
  booktitle={Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval},
  pages={378--387},
  year={2021}
}

The code is tested under a Linux desktop with TensorFlow 1.15.2 and Python 3.6.8.

Data Pre-processing

The script is reco_utils/dataset/sequential_reviews.py which will be automatically excuted when there exists no pre-processed training file.

Model Training

To train our model on Kuaishou dataset (with default hyper-parameters):

python examples/00_quick_start/sequential.py --dataset kuaishou

or on Taobao dataset:

python examples/00_quick_start/sequential.py --dataset taobao

Misc

The implemention is based on Microsoft Recommender.