How To Break Anonymity of the Netflix Prize Dataset

Abstract

We present a new class of statistical de-anonymization attacks against high-dimensional micro-data, such as individual preferences, recommendations, transaction records and so on. Our techniques are robust to perturbation in the data and tolerate some mistakes in the adversary's background knowledge. We apply our de-anonymization methodology to the Netflix Prize dataset, which contains anonymous movie ratings of 500,000 subscribers of Netflix, the world's largest online movie rental service. We demonstrate that an adversary who knows only a little bit about an individual subscriber can easily identify this subscriber's record in the dataset. Using the Internet Movie Database as the source of background knowledge, we successfully identified the Netflix records of known users, uncovering their apparent political preferences and other potentially sensitive information.

Cited by in corpus (26)

• Differentially Private Distributed Constrained Optimization
• On Lightweight Privacy-Preserving Collaborative Learning for IoT Objects
• Mitigating Manipulation in Peer Review via Randomized Reviewer Assignments
• Using Blockchain to Achieve Decentralized Privacy In IoT Healthcare
• DP-InstaHide: Provably Defusing Poisoning and Backdoor Attacks with Differentially Private Data Augmentations
• Scalable Realistic Recommendation Datasets through Fractal Expansions
• Anonymizing Machine Learning Models
• Differential Privacy Made Easy
• Privacy-Preserving Blockchain Based Federated Learning with Differential Data Sharing
• Privacy-Aware Recommender Systems Challenge on Twitter's Home Timeline
• Multi-Resolution Diffusion for Privacy-Sensitive Recommender Systems
• Improved Differentially Private Decentralized Source Separation for fMRI Data
• DALock: Distribution Aware Password Throttling
• Exactly Optimal and Communication-Efficient Private Estimation via Block Designs
• Differentially Private Convex Optimization with Piecewise Affine Objectives
• Scaling Up Collaborative Filtering Data Sets through Randomized Fractal Expansions
• Power to the Data Defenders: Human-Centered Disclosure Risk Calibration of Open Data
• Noise Variance Optimization in Differential Privacy: A Game-Theoretic Approach Through Per-Instance Differential Privacy
• Design of Privacy-Preserving Dynamic Controllers
• A Framework for Private Matrix Analysis
• Learning Numeric Optimal Differentially Private Truncated Additive Mechanisms
• Nebula: Efficient, Private and Accurate Histogram Estimation
• Privately Learning Subspaces
• Corella: A Private Multi Server Learning Approach based on Correlated Queries
• The 2021 RecSys Challenge Dataset: Fairness is not optional
• Privacy Parameter Variation Using RAPPOR on a Malware Dataset