网络安全与数据治理2025，Vol.44Issue(11)：7-11,17,6.DOI:10.19358/j.issn.2097-1788.2025.11.002
基于同态加密的AI模型参数安全计算与防泄露方法
Secure computation and anti-leakage methods for AI model parameters based on homomorphic encryption
张恒 1廖尚斌 1张陈颖1
作者信息
- 1. 中国移动通信集团福建有限公司,福建 福州 350000
- 折叠
摘要
Abstract
With the extensive application of artificial intelligence in sensitive fields such as healthcare and finance,the privacy protection of model parameters and training data has become a critical issue.This paper proposes a secure computation and anti-leakage method for AI model parameters based on homomorphic encryption(HE).The method employs the CKKS scheme to im-plement parameter encryption,forward inference,and gradient updates in the ciphertext space,thereby avoiding the risk of plain-text exposure during training.The results demonstrate that HE-SGD achieves a maximum accuracy of 99.1%on MNIST.In terms of computational overhead,it balances efficiency and security,with an information leakage risk index close to 0.0.The study in-dicates that the proposed method maintains model precision while achieving efficient and secure computation with nearly zero leak-age risk,showing strong application value.关键词
模型参数/隐私保护/同态加密/CKKS方案/梯度更新Key words
model parameters/privacy protection/homomorphic encryption/CKKS scheme/gradient updates分类
计算机与自动化引用本文复制引用
张恒,廖尚斌,张陈颖..基于同态加密的AI模型参数安全计算与防泄露方法[J].网络安全与数据治理,2025,44(11):7-11,17,6.
