Cryptology ePrint Archive: Report 2022/322

SecFloat: Accurate Floating-Point meets Secure 2-Party Computation

Deevashwer Rathee and Anwesh Bhattacharya and Rahul Sharma and Divya Gupta and Nishanth Chandran and Aseem Rastogi

Abstract: We build a library SecFloat for secure 2-party computation (2PC) of 32-bit single-precision floating-point operations and math functions. The existing functionalities used in cryptographic works are imprecise and the precise functionalities used in standard libraries are not crypto-friendly, i.e., they use operations that are cheap on CPUs but have exorbitant cost in 2PC. SecFloat bridges this gap with its novel crypto-friendly precise functionalities. Compared to the prior cryptographic libraries, SecFloat is up to six orders of magnitude more precise and up to two orders of magnitude more efficient. Furthermore, against a precise 2PC baseline, SecFloat is three orders of magnitude more efficient. The high precision of SecFloat leads to the first accurate implementation of secure inference. All prior works on secure inference of deep neural networks rely on ad hoc float-to-fixed converters. We evaluate a model where the fixed-point approximations used in privacy-preserving machine learning completely fail and floating-point is necessary. Thus, emphasizing the need for libraries like SecFloat.

Category / Keywords: cryptographic protocols / secure two-party computation; floating-point; privacy-preserving machine learning; secure inference; privacy-preserving proximity testing;

Original Publication (with major differences): IEEE Security and Privacy 2022

Date: received 8 Mar 2022, last revised 8 Mar 2022

Contact author: deevashwer at berkeley edu, rahsha at microsoft com, divya gupta at microsoft com

Available format(s): PDF | BibTeX Citation

Version: 20220308:125129 (All versions of this report)

Short URL: ia.cr/2022/322


[ Cryptology ePrint archive ]