Multi-Selection for Recommendation Systems
Preprint
Abstract
We present the construction of a multi-selection model to answer differentially private queries in the context of recommendation systems. The server sends back multiple recommendations and a ``local model'' to the user, which the user can run locally on its device to select the item that best fits its private features. We study a setup where the server uses a deep neural network (trained on the Movielens 25M dataset as the ground truth for movie recommendation. In the multi-selection paradigm, the average recommendation utility is approximately 97\% of the optimal utility (as determined by the ground truth neural network) while maintaining a local differential privacy guarantee with $ε$ ranging around 1 with respect to feature vectors of neighboring users. This is in comparison to an average recommendation utility of 91\% in the non-multi-selection regime under the same constraints.
Keywords
All Related Versions
This publication has 16 references indexed in Scilit:
- Private and Utility Enhanced Recommendations With Local Differential Privacy and Gaussian Mixture ModelIEEE Transactions on Knowledge and Data Engineering, 2021
- Privacy Enhanced Matrix Factorization for Recommendation with Local Differential PrivacyIEEE Transactions on Knowledge and Data Engineering, 2018
- Interpretable Explanations of Black Boxes by Meaningful Perturbation2017
- The MovieLens DatasetsACM Transactions on Interactive Intelligent Systems, 2015
- RAPPOR2014
- Differential Privacy: An Economic Method for Choosing Epsilon2014
- Local Privacy and Statistical Minimax Rates2013
- Private traits and attributes are predictable from digital records of human behaviorProceedings of the National Academy of Sciences, 2013
- Geo-indistinguishability2013
- Mechanism Design via Differential Privacy2007