UniCDR
The source code is for the paper: “Towards Universal Cross-Domain Recommendation” accepted in WSDM 2023 by Jiangxia Cao, Shaoshuai Li, Bowen Yu, Xiaobo Guo, Tingwen Liu and Bin Wang.
@inproceedings{cao2023unicdr,
title={Towards Universal Cross-Domain Recommendation},
author={Cao, Jiangxia and Li, Shaoshuai and Yu, Bowen and Guo, Xiaobo and Liu, Tingwen and Wang, Bin},
booktitle={ACM International Conference on Web Search and Data Mining (WSDM)},
year={2023}
}
Requirements
Python=3.7.9
PyTorch=1.6.0
Scipy = 1.5.2
Numpy = 1.19.1
Datasets
We release a large-scale dataset for multi-domain recommendation in this work, and the all used datasets can be downloaded at WSDM2023-UniCDR-datasets (including 4 CDR scenarios: dual-user-intra, dual-user-inter, multi-item-intra, and multi-user-intra).
Note that all datasets are required to unzip in the root directory.
Usage
To run this project, please make sure that you have the following packages and datasets being downloaded. Our experiments are conducted on a PC with an Intel Xeon E5 2.1GHz CPU, 256 RAM and a Tesla T4 16GB GPU.
Running example:
# dual-user-intra
CUDA_VISIBLE_DEVICES=0 python -u train_rec.py --static_sample --cuda --domains sport_cloth --aggregator user_attention > dual_user_intra_sport_cloth.log 2>&1&
# dual-user-inter
CUDA_VISIBLE_DEVICES=1 python -u train_rec.py --static_sample --cuda --domains game_video --task dual-user-inter --aggregator mean > dual_item_inter_game_video.log 2>&1&
# multi-item-intra
CUDA_VISIBLE_DEVICES=2 python -u train_rec.py --static_sample --cuda --domains m1_m2_m3_m4_m5 --task multi-item-intra --aggregator item_similarity > multi_item_intra.log 2>&1&
# multi-user-intra
CUDA_VISIBLE_DEVICES=3 python -u train_rec.py --static_sample --cuda --domains d1_d2_d3 --task multi-user-intra --aggregator mean > multi_user_intra.log 2>&1&