Title:PrivateRec: Differentially Private Training and Serving for Federated News Recommendation

Abstract:Collecting and training over sensitive personal data raise severe privacy concerns in personalized recommendation systems, and federated learning can potentially alleviate the problem by training models over decentralized user this http URL, a theoretically private solution in both the training and serving stages of federated recommendation is essential but still this http URL, naively applying differential privacy (DP) to the two stages in federated recommendation would fail to achieve a satisfactory trade-off between privacy and utility due to the high-dimensional characteristics of model gradients and hidden this http URL this work, we propose a federated news recommendation method for achieving a better utility in model training and online serving under a DP this http URL first clarify the DP definition over behavior data for each round in the life-circle of federated recommendation this http URL, we propose a privacy-preserving online serving mechanism under this definition based on the idea of decomposing user embeddings with public basic vectors and perturbing the lower-dimensional combination coefficients. We apply a random behavior padding mechanism to reduce the required noise intensity for better utility. Besides, we design a federated recommendation model training method, which can generate effective and public basic vectors for serving while providing DP for training participants. We avoid the dimension-dependent noise for large models via label permutation and differentially private attention modules. Experiments on real-world news recommendation datasets validate that our method achieves superior utility under a DP guarantee in both training and serving of federated news recommendations.