Abstract:Modern recommender systems are trained to predict users potential future interactions from users historical behavior data. During the interaction process, despite the data coming from the user side recommender systems also generate exposure data to provide users with personalized recommendation slates. Compared with the sparse user behavior data, the system exposure data is much larger in volume since only very few exposed items would be clicked by the user. Besides, the users historical behavior data is privacy sensitive and is commonly protected with careful access authorization. However, the large volume of recommender exposure data usually receives less attention and could be accessed within a relatively larger scope of various information seekers. In this paper, we investigate the problem of user behavior leakage in recommender systems. We show that the privacy sensitive user past behavior data can be inferred through the modeling of system exposure. Besides, one can infer which items the user have clicked just from the observation of current system exposure for this user. Given the fact that system exposure data could be widely accessed from a relatively larger scope, we believe that the user past behavior privacy has a high risk of leakage in recommender systems. More precisely, we conduct an attack model whose input is the current recommended item slate (i.e., system exposure) for the user while the output is the user's historical behavior. Experimental results on two real-world datasets indicate a great danger of user behavior leakage. To address the risk, we propose a two-stage privacy-protection mechanism which firstly selects a subset of items from the exposure slate and then replaces the selected items with uniform or popularity-based exposure. Experimental evaluation reveals a trade-off effect between the recommendation accuracy and the privacy disclosure risk.

Research on Recommender System Model Based on Differential Privacy and Time Series

When Differential Privacy Meets Randomized Perturbation: A Hybrid Approach For Privacy-Preserving Recommender System

Towards Differential Privacy in Sequential Recommendation: A Noisy Graph Neural Network Approach

Differential privacy in collaborative filtering recommender systems: a review

The Movie Recommendation System Based on Differential Privacy

Studying the Impact of Data Disclosure Mechanism in Recommender Systems via Simulation

Model Stealing Attack against Recommender System

A GCN-based Model for Recommendation Using Local Differential Privacy Method

A Novel Differential Privacy Protection Model for Publishing Time Series of Classification Task

SDRM-LDP: A Recommendation Model Based on Local Differential Privacy

A differentially private matrix factorization based on vector perturbation for recommender system

A differentially private nonnegative matrix factorization for recommender system

An Effective Federated Recommendation Framework with Differential Privacy

On the User Behavior Leakage from Recommender System Exposure

DPLCF: Differentially Private Local Collaborative Filtering

Differential Data Analysis for Recommender Systems

Position Paper on Simulating Privacy Dynamics in Recommender Systems

Probabilistic Matrix Factorization with Personalized Differential Privacy

An efficient privacy-preserving point-of-interest recommendation model based on local differential privacy

Temporal‐aware and sparsity‐tolerant hybrid collaborative recommendation method with privacy preservation