Matrix Factorization for Practical Continual Mean Estimation Under User-Level Differential Privacy

ArXi:2601.22320v2 Announce Type: replace We study continual mean estimation, where data vectors arrive sequentially and the goal is to maintain accurate estimates of the running mean. We address this problem under user-level differential privacy, which protects each user's entire dataset even when they contribute multiple data points. Previous work on this problem has focused on pure differential privacy. While important, this approach limits applicability, as it leads to overly noisy estimates.