摘要
神经网络数据传递方式不断多元化,使得数据存在泄漏及被攻击的风险,通过对抗加密算法实现神经网络模型的改进,能够优化网络系统抵御风险的能力,保证神经网络模型使用的安全性。本研究首先分析了对抗神经网络数据密码攻击方式及密码安全检测技术,在此基础上进行对抗加密算法模型的改进设计,得到CCA-ANC模型,并通过仿真实验对改进后的模型设计进行验证,最终得出神经网络模型具有安全高效的应用效果,对于后期推广应用具有一定的借鉴价值。
Due to the continuous diversification of neural network data transmission methods,there are risks of data leakage and attack.The improvement of neural network model through anti-encryption algorithm can optimize the ability of network system to resist risks and ensure the security of neural network model.Firstly,this research analyzes the cryptographic attack mode and cryptographic security detection technology of the anti-neural network data.On this basis,the improved design of the anti-cryptographic algorithm model is carried out to obtain the CCA-ANC model,and the improved model design is verified by simulation experiment.Finally,it is concluded that the neural network model has a secure and efficient application effect.It has certain reference value for later popularization and application.
作者
陈其宇
许海兰
Chen Qiyu;Xu Hailan(Shenzhen Suoxinda Data Technology Co.,Ltd.,Shenzhen,China)
出处
《科学技术创新》
2023年第8期108-111,共4页
Scientific and Technological Innovation
关键词
对抗加密算法
神经网络
数据加密技术
仿真验算
adversarial encryption algorithm
neural network
data encryption technology
simulation checking
