A Huber Loss Minimization Approach to Mean Estimation under User-level Differential Privacy Puning Zhao Lifeng Lai Li Shen Zhejiang Lab University of California, Davis Sun Yat-Sen University

Privacy protection of users' entire contribution of samples is important in distributed systems. The most effective approach is the two-stage scheme, which finds a small interval first and then gets a refined estimate by clipping samples into the interval. However, the clipping operation induces bias, which is serious if the sample distribution is heavy-tailed. Besides, users with large local sample sizes can make the sensitivity much larger, thus the method is not suitable for imbalanced users. Motivated by these challenges, we propose a Huber loss minimization approach to mean estimation under user-level differential privacy.