HOMOMORPHIC ENCRYPTION-BASED PRIVACY PRESERVATION USING DATA SHARING IN CLOUD ENVIRONMENTS
DOI:
https://doi.org/10.5281/zenodo.21130485Abstract
Decentralized and collaborative learning approach that ensures the data privacy of each participant. However, recent studies have shown that the private data of each participant can be obtained from shared parameters of local models through reversal model and membership inference attacks leading to privacy leakage. Privacy-preserving federated learning strategies based on Homomorphic Encryption (PPFL-HE) have been developed to solve this issue. PPFL-HE methods require high communication and computational overheads, which are impractical for resource-limited devices. This work proposes an efficient PPFL-HE method to reduce communication and computational overheads. The suggested approach is predicated on a novel quantization procedure that encodes the weights of the local models into long-signed integers by introducing a dynamic range evaluation layer-for-layer (DREL). The suggested method lowers the quantization mistakes and communication overhead in comparison to conventional quantization techniques. Additionally, it makes it possible to encrypt local weights using the Brakerski/Fan-Vercauteren Homomorphic Encryption method (BFVHE), which is very effective on integers and reduces ciphertext size, encryption, aggregation, and decryption time. Compared to other PPFL-HE techniques, the suggested approach has lower computing overheads and communication costs. In particular, the suggested approach reduces encryption on average when compared to PPFL-HE approaches based on Paillier and CKKS schemes, respectively, and necessitates global model convergence.
Downloads
Published
Issue
Section
License

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.













