Speaker: 郑晓鹏讲师，汕头大学

Inviter: 王定康研究员

Title: A New Framework for Fast Homomorphic Matrix Multiplication

Language: Chinese

Time & Venue: 2024.12.09 15:00-16:00 腾讯会议：697-640-246

Abstract: Homomorphic Encryption (HE) is one of the mainstream cryptographic tools used to enable secure outsourced computation. A typical task is secure matrix computation, which is a fundamental operation used in various outsourced computing applications such as statistical analysis and machine learning. In this paper, we present a new framework for secure multiplication of two matrices with size r×s and s ×t respectively, which requires only O(logs) basic homomorphic operations, whereas the state-of-the-art methods require at least O(s) basic homomorphic operations. We implemented our method together with other methods using the BGV scheme in HElib. Experimental results show that the new framework has significant advantage in efficiency when r × s × t ≤ n, where n is dimension of the polynomial ring used in RLWE encryption. In this case, the new framework is 4.0 to 106.8 times faster than exiting algorithms in experiments.