ByzSecAgg: A Byzantine-Resistant Secure Aggregation Scheme for Federated Learning Based on Coded Computing and Vector Commitment. (arXiv:2302.09913v2 [cs.CR] UPDATED)

In this paper, we propose an efficient secure aggregation scheme for
federated learning that is protected against Byzantine attacks and privacy
leakages. Processing individual updates to manage adversarial behavior, while
preserving privacy of data against colluding nodes, requires some sort of
secure secret sharing. However, communication load for secret sharing of long
vectors of updates can be very high. To resolve this issue, in the proposed
scheme, local updates are partitioned into smaller sub-vectors and shared using
ramp secret sharing. …

adversarial aggregation attacks computing data federated learning manage nodes privacy sort updates