## Cryptology ePrint Archive: Report 2020/451

Maliciously Secure Matrix Multiplication with Applications to Private Deep Learning

Hao Chen and Miran Kim and Ilya Razenshteyn and Dragos Rotaru and Yongsoo Song and Sameer Wagh

Abstract: Computing on data in a manner that preserve the privacy is of growing importance. Secure Multi-Party Computation (MPC) and Homomorphic Encryption (HE) are two cryptographic techniques for privacy-preserving computations. In this work, we have developed efficient UC-secure multiparty protocols for matrix multiplications and two-dimensional convolutions. We built upon the SPDZ framework and integrated the state-of-the-art HE algorithms for matrix multiplication. We also optimized the zero-knowledge proofs and the sacrifice'' step of SPDZ to further improve efficiency. As a result, our protocol achieved communication cost linear only on the input and output dimensions and not on the number of multiplication operations. We implemented our protocols and benchmarked them against the SPDZ LowGear variant (Keller et al. Eurocrypt'18). For multiplying two square matrices of size 128, we reduced the communication cost from 1.54 GB to 12.46 MB, an improvement of over two orders of magnitude that only improves with larger matrix sizes. For evaluating all convolution layers of the ResNet-50 neural network, we reduced the communication cost from 5 TB to 41 GB.

Category / Keywords: cryptographic protocols / cryptographic protocols / Multiparty computation, Dishonest majority, Homomorphic encryption

Date: received 17 Apr 2020, last revised 18 Apr 2020

Contact author: swagh at princeton edu,haoche@microsoft com,miran kim@uth tmc edu,ilyaraz@microsoft com,dragos rotaru@esat kuleuven be,yongsoo song@microsoft com

Available format(s): PDF | BibTeX Citation

Short URL: ia.cr/2020/451

[ Cryptology ePrint archive ]