We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CR

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Cryptography and Security

Title: Efficient Secure Aggregation Based on SHPRG For Federated Learning

Abstract: We propose a novel secure aggregation scheme based on seed-homomorphic pseudo-random generator (SHPRG) to prevent private training data leakage from model-related information in Federated Learning systems. Our constructions leverage the homomorphic property of SHPRG to simplify the masking and demasking scheme, which entails a linear overhead while revealing nothing beyond the aggregation result against colluding entities. Additionally, our scheme is resilient to dropouts without extra overhead. We experimentally demonstrate our scheme significantly improves the efficiency to 20 times over baseline, especially in the more realistic case in which the number of clients and model size become large and a certain percentage of clients drop out from the system.
Comments: 6 pages, 3 figures
Subjects: Cryptography and Security (cs.CR); Distributed, Parallel, and Cluster Computing (cs.DC)
Cite as: arXiv:2111.12321 [cs.CR]
  (or arXiv:2111.12321v1 [cs.CR] for this version)

Submission history

From: Zizhen Liu [view email]
[v1] Wed, 24 Nov 2021 08:08:39 GMT (490kb)

Link back to: arXiv, form interface, contact.