Score: 2

Multi-Selection for Recommendation Systems

Published: April 10, 2025 | arXiv ID: 2504.07403v1

By: Sahasrajit Sarmasarkar , Zhihao Jiang , Ashish Goel and more

BigTech Affiliations: Stanford University Princeton University

Potential Business Impact:

Keeps your movie picks private, still good.

Business Areas:
A/B Testing Data and Analytics

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 $\epsilon$ 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.

Country of Origin
🇺🇸 United States

Page Count
22 pages

Category
Computer Science:
Machine Learning (CS)