基于小样本学习的网络异常流量检测  

Abnormal Network Traffic Detection Based on Small Sample Learning

作　　者：李荣宽 丁乙 王寒凝 贺宁 LI Rongkuan;DING Yi;WANG Hanning;HE Ning(CETC Cloud(Beijing)Technology Co.Ltd.,Beijing 100041,China;Unit 61932 of PLA,Beijing 100000,China;School of Cyber Science and Engineering,Southeast University,Nanjing 211189,China)

机构地区：[1]电科云(北京)科技有限公司,北京100041 [2]解放军61932部队,北京100000 [3]东南大学网络空间安全学院,南京211189

出　　处：《指挥信息系统与技术》2024年第2期88-93,共6页Command Information System and Technology

基　　金：军委科技委基金资助项目。

摘　　要：网络结构具有较高复杂性,因此导致各种异常流量现象层出不穷,其中包括一些标注样本极少的新型异常流量类型。为了有效识别标注样本量极少的异常情况,设计了一种基于小样本学习的网络异常流量检测方法。该方法利用基于小样本的迁移学习技术识别异常流量,从而确保了网络安全。The high complexity of network structure leads to various abnormal traffic phenomena,including some new abnormal traffic types with few labeled samples.In order to effectively identify the abnormal situations with few labeled samples,an abnormal network traffic detection method based on small sample learning is designed.The method uses the transfer learning technology based on small samples to identify the abnormal traffic.Thus,it can ensure the network security.

关 键 词：小样本 迁移学习 网络异常流量 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]