A parameter-free Bayesian aggregation method, based on variational inference over latent honesty indicators, matches Krum's robustness on benchmark federated learning attacks without needing to know the number of malicious clients.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Bayesian Robust Aggregation for Federated Learning
A parameter-free Bayesian aggregation method, based on variational inference over latent honesty indicators, matches Krum's robustness on benchmark federated learning attacks without needing to know the number of malicious clients.