New framework boosts security and efficiency for federated learning

Researchers have developed a new framework for federated learning that enhances security and efficiency for sign-based methods. This approach ensures information-theoretic security by securely computing the majority vote polynomial, revealing only the aggregated sign to the server. The framework incorporates techniques like inverse-form exponent reduction and single-round secure multiplication to significantly decrease communication and latency, achieving up to 99.5% reduction in online communication. Additionally, it offers robustness against dropouts and adversarial behaviors, leading to substantial accuracy gains. AI

IMPACT Enhances privacy and efficiency in on-device federated learning, potentially enabling wider adoption for resource-constrained devices.

RANK_REASON The item is a research paper published on arXiv detailing a new technical framework for federated learning. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• alphaXiv
• arXiv
• CatalyzeX
• CORE Recommender
• DagsHub
• federated learning
• Gotit.pub
• Hugging Face
• IArxiv Recommender
• MDS code
• MV polynomial
• ScienceCast
• signSGD

• paper
• safety
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →