Testing hybrid homomorphic encryption inside federated learning on MNIST shows client uploads drop 16x versus full homomorphic encryption with identical accuracy, while server-side transciphering becomes the new bottleneck.
In: International conference on security and privacy in communication systems
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
A Privacy-Centric Approach: Scalable and Secure Federated Learning Enabled by Hybrid Homomorphic Encryption
Testing hybrid homomorphic encryption inside federated learning on MNIST shows client uploads drop 16x versus full homomorphic encryption with identical accuracy, while server-side transciphering becomes the new bottleneck.