首页|面向智能无人通信系统的因果性对抗攻击生成算法

面向智能无人通信系统的因果性对抗攻击生成算法

扫码查看
原文链接
  • 万方数据
  • 维普
考虑到基于梯度的对抗攻击生成算法在实际通信系统部署中面临着因果性问题,提出了一种因果性对抗攻击生成算法.利用长短期记忆网络的序列输入输出特征与时序记忆能力,在满足实际应用中存在的因果性约束前提下,有效提取通信信号的时序相关性,增强针对无人通信系统的对抗攻击性能.仿真结果表明,所提算法在同等条件下的攻击性能优于泛用对抗扰动等现有的因果性对抗攻击生成算法.
Causality adversarial attack generation algorithm for intelligent unmanned communication system
A causality adversarial attack generation algorithm was proposed in response to the causality issue of gradi-ent-based adversarial attack generation algorithms in practical communication system.The sequential input-output fea-tures and temporal memory capability of long short-term memory networks were utilized to extract the temporal correla-tion of communication signals while satisfying practical causality constraints,and enhance the adversarial attack perfor-mance against unmanned communication systems.Simulation results demonstrate that the proposed algorithm outper-forms existing causality adversarial attack algorithms,such as universal adversarial perturbation,under identical conditions.

intelligent communication systemadversarial attackdeep learningcausal systemlong short-term memory network

禹树文、许威、姚嘉铖

展开 >

东南大学移动通信全国重点实验室,江苏 南京 210096

网络通信与安全紫金山实验室,江苏 南京 211111

智能通信系统 对抗攻击 深度学习 因果系统 长短期记忆网络

国家自然科学基金资助项目国家自然科学基金资助项目中央高校基本科研业务费专项资金资助项目中央高校基本科研业务费专项资金资助项目

62022026622115301082242022K600022242023K5003

2024

10.11959/j.issn.1000-436x.2024036

通信学报
中国通信学会

通信学报

CSTPCD北大核心
影响因子：1.265
ISSN：1000-436X
年,卷(期)：2024.45(1)
  • 25