基于迁移学习的网络身份双向认证技术

Network Identity Bidirectional Authentication Technology Based on Transfer Learning

现有网络身份认证技术的算法复杂性高,网络身份认证技术的同步性较差,因此提出一种基于迁移学习的网络身份双向认证技术。将目标分类器和相似性约束相联系,通过迁移学习以及相关领域弱相似性,训练用户身份数据,针对加密机制中特征2域的二进制更容易在系统中实现的优点,确定出网络身份认证过程中曲线和基点,以二者为依据分别在读写器和标签两个方向设置公钥和私钥,并在每一次认证后更新一次基点,使私钥序列不断改变,实现网络身份双向认证。通过实验表明,所提方法的同步性强,而且能够拦截多种网络攻击。

Because of the complexity of the existing network identity authentication technology,the synchronization of network identity authentication technology is poor.A two-way authentication technology based on migration learning is proposed.By linking the target classifier with similarity constraints,training user identity data through migration learning and weak similarity in related fields,and aiming at the advantages that the binary of feature 2 domain in encryption system is easier to realize in the system,the curve and base point in the process of network identity authentication are determined.The public key and the private key are set in the two directions of reader and label,and the base point is updated after each authentication,which makes the private key sequence change continuously,realizes network identity two-way authentication.Experiments show that the proposed method has strong synchronization and can intercept multiple network attacks.