Abstract
This talk aims to discuss the case when data are distributed and different differential privacy constraints are imposed on each of the servers/users. Across servers/users, only private information is allowed to be transmitted. Different classical statistical inference problems will be discussed, accompanied with their fundamental limits.