Skip to content

Latest commit

 

History

History
29 lines (23 loc) · 908 Bytes

README.md

File metadata and controls

29 lines (23 loc) · 908 Bytes

Single-User Injection for Invisible Shilling Attack against Recommender Systems

A PyTorch implementation of paper:

Single-User Injection for Invisible Shilling Attack against Recommender Systems, Chengzhi Huang, Hui Li , CIKM '2023

Requirements

dgl==0.4.3
numpy>=1.15
pandas>=0.19
scipy>=0.18
torch>=1.3
higher

Floder

  • models contains influence module, recommender and baseline attacker
  • config contains its super parameters

Data

The dataset used in experiments or other scripts can be download from Google Drive

How to run

python main.py

other examples can be shown in jupyter notebook.

The parameters args.do_train and args.do_eval control whether the program is training the model or evaluating the model. More details can be found in the paper.