m1234543 / arlib

An open-source framework for conducting data poisoning attacks on recommendation systems, designed to assist researchers and practitioners.

ARLib (Pytorch)

An open-source framework for conducting data poisoning attacks on recommendation systems, designed to assist researchers and practitioners.

Members:
Zongwei Wang, Chongqing University, China, [email protected]
Hao Ma, Chongqing University, China, [email protected]

Supported by:
Prof. Min Gao, Chongqing University, China, [email protected]

Framework

Usage

1. Two configure files attack_parser.py and recommend_parser are in the directory named conf, and you can select and configure the recommendation model and attack model by modifying the configuration files.
   
2. Run main.py.

Implemented Models

Recommend Model	Paper	Type
GMF	Yehuda et al. Matrix Factorization Techniques for Recommender Systems, IEEE Computer'09.	MF
WRMF	Hu et al.Collaborative Filtering for Implicit Feedback Datasets, KDD'09.	MF
NCF	He et al. Neural Collaborative Filtering, WWW'17.	Deep Learning
NGCF	Wang et al. Neural Graph Collaborative Filtering, SIGIR'19.	Graph
SGL	Wu et al. Self-supervised Graph Learning for Recommendation, SIGIR'21.	Graph + CL
SimGCL	Yu et al. Are Graph Augmentations Necessary? Simple Graph Contrastive Learning for Recommendation, SIGIR'22.	Graph + CL

• CL is short for contrastive learning (including data augmentation); DA is short for data augmentation only

Attack Model	Paper	Form	Method
RandomAttack	Lam et al. Shilling Recommender Systems for Fun and Profit. WWW'2004	dataAttack	Heuristic
BandwagonAttack	Gunes et al. Shilling Attacks against Recommender Systems: A Comprehensive Survey. Artif.Intell.Rev.'2014	dataAttack	Heuristic
PGA	Li et al. Data poisoning attacks on factorization-based collaborative filtering. NIPS'2016.	dataAttack	Direct Gradient Optimization
AUSH	Lin C et al. Attacking recommender systems with augmented user profiles. CIKM'2020.	dataAttack	GAN
GOAT	Wu et al. Ready for emerging threats to recommender systems? A graph convolution-based generative shilling attack. Information Sciences'2021.	dataAttack	GAN
FedRecAttack	Rong et al. Fedrecattack: Model poisoning attack to federated recommendation. ICDE'2022.	gradientAttack	Direct Gradient Optimization
RAPU_G	Zhang et al. Data Poisoning Attack against Recommender System Using Incomplete and Perturbed Data. KDD'2021).	dataAttack	optimization
A_ra	Rong et al. Poisoning Deep Learning Based Recommender Model in Federated Learning Scenarios. IJCAI'2022.	gradientAttack	Direct Gradient Optimization

Implement Your Model

Determine whether you want to implement the attack model or the recommendation model, and then add the file under the corresponding directory.

If you are an attack method, make sure：

1. Whether you need information of the recommender model, and then set self.recommenderGradientRequired.
   
2. Whether you need gradient information of training recommender model, and then set self.recommenderModelRequired.
   
3. Make sure your attack type (gradientAttack/dataAttack).
   

• If gradientAttack: Reimplement function gradientattack()
  
• If dataAttack: Reimplement function gradientattack()
  

If you are an attack method, reimplement the following functions：

• init()
• train()
• save()
• predict()
• evaluate()
• test()

Requirements

base==1.0.4
numba==0.53.1
numpy==1.18.0
scipy==1.4.1
torch==1.7.1