grcpeng/dccl

Hi, welcome to our DCCL (Disentangled Causal Embedding With Contrastive Learning For Recommender System. Accepted by the industry track of WWW'23) program.

KS-DCCL

This code was used for offline experiments of DCCL. You can reproduce the offline experiments in the paper by executing

python main.py

KS-dataset

The industrial Short-video dataset is collected from Kuaishou (a billion-user scale short-video recommender system: https://www.kuaishou.com/new-reco). However, due to national policy restrictions, our database cannot be released at present, and we apologize for this.

Cite

Please cite the associated paper for this work if you use this code:

@article{zhao2023disentangled,
  title={Disentangled Causal Embedding With Contrastive Learning For Recommender System},
  author={Zhao, Weiqi and Tang, Dian and Chen, Xin and Lv, Dawei and Ou, Daoli and Li, Biao and Jiang, Peng and Gai, Kun},
  journal={arXiv preprint arXiv:2302.03248},
  year={2023}
}