Search papers, labs, and topics across Lattice.
This paper introduces a lightweight, information-theoretically secure aggregation framework for sign-based federated learning (FL), addressing vulnerabilities in gradient sign exposure while minimizing communication costs. By employing single-round secure multiplication and inverse-form exponent reduction, the framework significantly reduces online communication by up to 99.5% and latency by up to 85.7%, all while maintaining robust performance against dropouts and adversaries. The proposed approach achieves accuracy improvements of up to 20.65% and 10.74% in the presence of dropout and adversarial attacks, respectively, establishing a new standard for secure FL implementations on resource-constrained devices.
Achieving up to 99.5% reduction in communication costs while ensuring information-theoretic security could revolutionize federated learning on edge devices.
On-device federated learning (FL) enables privacy-preserving and personalized model training on resource-constrained devices such as smartphones and IoT nodes. To reduce communication cost, sign-based methods (e.g., signSGD) transmit one-bit gradients. However, exposing gradient signs makes them vulnerable to inference attacks, while existing secure aggregation schemes are often incompatible with such methods or incur significant computational and communication overhead. We propose a lightweight and information-theoretically secure aggregation framework tailored for sign-based FL. The framework securely computes the majority vote (MV) polynomial through single-round secure multiplication, ensuring end-to-end information-theoretic security under the honest-majority assumption while revealing only the final aggregated sign to the server. To enhance efficiency and scalability, we introduce two key techniques. First, inverse-form exponent reduction halves the effective MV polynomial degree, reducing both communication and computation costs. Second, we propose single-round secure multiplication, achieving linear offline complexity and storage with only a single online communication. Together, these techniques reduce online communication by up to 99.5% and latency by up to 85.7% compared to conventional approaches. Also, by leveraging inherent MDS-code-based decoding, the framework achieves robustness against both dropouts and adversarial behaviors, yielding accuracy gains of up to 20.65% and 10.74%, respectively. Overall, the proposed framework establishes a practical foundation for large-scale, low-latency, and information-theoretically secure aggregation in sign-based FL.
