This repository contains our implementations for Shilling Attacks against Recommender Systems.

Folder structure:

• AUSH: The implementation of AUSH used in our CIKM'20 paper [ACM Library] [arXiv Preprint].
• Leg-UP: The implementation of Leg-UP in our TNNLS'22 paper [IEEE Xplore] [arXiv Preprint] and a unified framework for comparing Leg-UP with various attackers including AIA, DCGAN, WGAN, Random Attack, Average Attack, Segment Attack and Bandwagon Attack.
• data: Recommendation datasets used in our experiments.

See README.md in each folder for more details.

Please kindly cite our papers if you find our implementations useful:

Chen Lin, Si Chen, Hui Li, Yanghua Xiao, Lianyun Li, and Qian Yang. 2020. Attacking Recommender Systems with Augmented User Profiles. In CIKM. 855–864.

Chen Lin, Si Chen, Meifang Zeng, Sheng Zhang, Min Gao, and Hui Li. 2022. Shilling Black-Box Recommender Systems by Learning to Generate Fake User Profiles. In TNNLS.

@inproceedings{Lin2020Attacking,  
  author    = {Chen Lin and
               Si Chen and
               Hui Li and
               Yanghua Xiao and
               Lianyun Li and
               Qian Yang},
  title     = {Attacking Recommender Systems with Augmented User Profiles},
  booktitle = {{CIKM}},
  pages     = {855--864},
  year      = {2020}
}  


@article{LinCZZGL22,
  author    = {Chen Lin and
               Si Chen and
               Meifang Zeng and
               Sheng Zhang and
               Min Gao and
               Hui Li},
  title     = {Shilling Black-Box Recommender Systems by Learning to Generate Fake User Profiles},
  journal   = {{IEEE} Trans. Neural Networks Learn. Syst.},
  year      = {2022}
}