ByITFL and LoByITFL are Byzantine-resilient federated learning schemes that claim perfect information-theoretic privacy using a polynomial discriminator and secret sharing, but the privacy proof has a gap around per-user norm validation.
Deep leakage from gradients,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Perfect Privacy for Discriminator-Based Byzantine-Resilient Federated Learning
ByITFL and LoByITFL are Byzantine-resilient federated learning schemes that claim perfect information-theoretic privacy using a polynomial discriminator and secret sharing, but the privacy proof has a gap around per-user norm validation.