Private Summation in the Shuffle Model
I will first introduce the shuffle model of differential privacy. Secondly I will present some forthcoming results, including a method for real valued summation in the one message shuffle model (where each party is limited to a single message), matching lower bounds on achievable accuracy and a new amplification by shuffling result for arbitrary local randomizers. Finally I shall present results on real summation in the many message shuffle model which follow from previous work on secure summation from anonymity by Ishai et al. and how to construct private summation from secure summation.