摘要
多用户语义通信,特别是异构语义网络是目前研究较少的课题。提出一种基于统一语义表征的单编码器-单解码器框架,其中异构的语义知识作为引导语义提取和表达的秘钥——符号消息被映射到统一的本原语义空间,并根据接收端用户各自的语义知识特异化解码。所提掩膜推理和异构语义转述约束,相较常用的点对点自编码方案,给语义可解释性及拓展性带来显著增强。针对多用户语义安全性,还提出一种利用语义知识流形空间进行语义加密的方法。
Multi-user semantic communication,typically heterogeneous semantic network,is a less-investigated topic.A singleencoder-single-decoder architecture based on unified semantic representation is proposed,under which heterogeneous semantic knowledge is utilized as the secret key that guides both semantic extraction and expression process-symbolic messages are first mapped into a unified primitive semantic space,and then decoded contingent on the receivers’own semantic knowledge.Compared with the commonly adopted point-to-point auto-encoding scheme,the proposed mask reasoning and heterogeneous semantic translation constraints significantly benefit the semantic interpretability and scalability.Targeted on multi-user semantic security,a semantic-level encryption method that manipulates the semantic knowledge manifold is further introduced.
作者
卢锟
李荣鹏
赵志峰
张宏纲
LU Kun;LI Rongpeng;ZHAO Zhifeng;ZHANG Honggang(College of Information Science and Electronic Engineering,Zhejiang University,Hangzhou 310027,China;Zhejiang Lab,Hangzhou 311121,China)
出处
《移动通信》
2023年第4期37-44,共8页
Mobile Communications
基金
国家自然科学基金(62071425)
浙江省“领雁”项目(2022C01093)
华为公司合作项目
浙江省杰出青年基金项目(LR23F010005)。
关键词
语义网络
语义表征
异构语义
多用户
semantic network
semantic representation
heterogeneous semantics
multi-user