Title:On the Information Theoretic Secure Aggregation with Uncoded Groupwise Keys

Abstract:Secure aggregation, which is a core component of federated learning, aggregates locally trained models from distributed users at a central server. The ``secure'' nature of such aggregation consists of the fact that no information about the local users' data must be leaked to the server except the aggregated local models. In order to guarantee security, some keys may be shared among the users. After the key sharing phase, each user masks its trained model which is then sent to the server. This paper follows the information theoretic secure aggregation problem originally formulated by Zhao and Sun, with the objective to characterize the minimum communication cost from the K users in the model aggregation phase. Due to user dropouts, the server may not receive all messages from the users. A secure aggregation scheme should tolerate the dropouts of at most $K-U$ users. The optimal communication cost is characterized by Zhao and Sun, but with the assumption that the keys stored by the users could be any random variables with arbitrary dependency. On the motivation that uncoded groupwise keys are more convenient to be shared and could be used in large range of applications besides federated learning, in this paper we assume the key variables are mutually independent and each key is shared by a group of S users. To the best of our knowledge, all existing secure aggregation schemes assign coded keys to the users. We show that if $S> K-U$, a new secure aggregation scheme with uncoded groupwise keys can achieve the same optimal communication cost as the best scheme with coded keys; if $S \leq K-U$, uncoded groupwise key sharing is strictly sub-optimal. Finally, we also implement our proposed secure aggregation scheme into Amazon EC2, which are then compared with the existing secure aggregation schemes with offline key sharing.