Answers are grounded in this paper and cite their sources.
We propose SwiftAgg+ , a novel secure aggregation protocol for federated learning systems, where a central server aggregates local models of <tex-math notation="LaTeX">N ∈ N </tex-math> distributed users, each of size <tex-math notation="LaTeX">L ∈ N </tex-math> , trained on their local data, in a privacy-preserving manner. SwiftAgg+ can significantly reduce the communication overheads without any compromise on security, and achieve optimal communication loads within diminishing gaps. Specifically, in presence of at most <tex-math notation="LaTeX">$D=o(N)$ </tex-math> dropout users, SwiftAgg+ achieves a per-user communication load of <tex-math notation="LaTeX">(1+O(1/N))L </tex-math> symbols and a server communication load of <tex-math notation="LaTeX">(1+O(1/N))L </tex-math> symbols, with a worst-case information-theoretic security guarantee, against any subset of up to <tex-math notation="LaTeX">$T=o(N)$ </tex-math> semi-honest users who may also collude with the curious server. Moreover, the proposed SwiftAgg+ allows for a flexible trade-off between communication loads and the number of active communication links. In particular, for <tex-math notation="LaTeX">$T< N-D$ </tex-math> and for any <tex-math notation="LaTeX">K∈ N </tex-math> , SwiftAgg+ can achieve the server communication load of <tex-math notation="LaTeX">(1+T/K)L </tex-math> symbols, and per-user communication load of up to <tex-math notation="LaTeX">(1+T+D/K)L </tex-math> symbols, where the number of pair-wise active connections in the network is <tex-math notation="LaTeX">N/2(K+T+D+1) </tex-math> .
No takes yet. Share an insight, caveat, or question.
Jahani-Nezhad et al. (2023) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: