Efficient Secure Matrix Multiplication Over LWE-Based Homomorphic Encryption

By Dung Hoang Duong, Pradeep Kumar Mishra, Masaya Yasuda. 2016

41 citations recorded by the source corpus.

Abstract Homomorphic encryption enables various calculations while preserving the data confidentiality. In this paper, we apply the somewhat homomorphic encryption scheme proposed by Brakerski and Vaikuntanathan (CRYPTO 2011) to secure matrix multiplication between two matrices. To reduce both the ciphertext size and the computation cost, we propose a new method to pack a matrix into a single ciphertexts so that it also enables efficient matrix multiplication over the packed ciphertexts. Our packing method generalizes Yasuda et al.’s methods (Security Comm. Networks 2015 and ACISP 2015), which are for secure inner product. We also implement our methods and give a comparison with previous packing methods.

Subfields: Artificial Intelligence, Computational Theory and Mathematics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Wei-Ting Chang
2. Ravi Tandon
3. Salim el Rouayheb

In Computational Theory and Mathematics

1. Ravi Tandon
2. Wei-Ting Chang

Citations in the corpus

• On the Capacity of Secure Distributed Matrix Multiplication (2018, background): “Secure matrix multiplication has been studied in cryptography community, and different approaches have been proposed, including a weaker version of fully homomorphic encryption, namely partially homomorphic encryption [6]–[8].”
• Secure Outsourced Matrix Computation and Application to Neural Networks (2018, unknown): “[18] proposed a method to encode a matrix as a constant polynomial in the native plaintext space.”