面向智能体的语义通信:架构与范例  

Semantic communication for intelligent devices:architectures and a paradigm

作　　者：张亦弛 张平[2] 魏急波[1] 赵海涛[1] 熊俊 张姣 Yichi ZHANG;Ping ZHANG;Jibo WEI;Haitao ZHAO;Jun XIONG;Jiao ZHANG(College of Electronic Science and Technology,National University of Defense Technology,Changsha 410073,China;State Key Laboratory of Networking and Switching Technology,Beijing University of Posts and Telecommunications,Beijing 100876,China)

机构地区：[1]国防科技大学电子科学学院,长沙410073 [2]北京邮电大学网络与交换技术国家重点实验室,北京100876

出　　处：《中国科学：信息科学》2022年第5期907-921,共15页Scientia Sinica(Informationis)

基　　金：国家自然科学基金(批准号:61931020,U19B2024,62001483);湖南省科技创新计划(批准号:2020RC2045)资助项目。

摘　　要：语义通信关注的是语义的准确传递,而不是数据或通信符号的准确传输,因此可以达到用更低的通信量实现更高效信息交互的目标,是适用于智能体交互的高效通信机制.本文分析总结了与语义通信紧密相关的一些研究工作,探讨了一种面向智能体的语义通信架构,并分析了该架构下各模块的功能;进一步,从有效性和可靠性两个方面分析了语义通信性能的指标和评估方法;最后,通过一个范例来说明语义通信的实现过程,证明其相较于传统通信方式可以显著降低通信量.Semantic communication focuses on the accurate transmission of meanings rather than data or symbols.It can achieve more efficient information interaction with less traffic,which is a promising communication mechanism for future intelligent devices.In this paper,we summarize existing research closely related to semantic communication.And then we propose a context-aware semantic communication architecture,explaining the functions of each module within it.We also introduce six metrics to evaluate the performance of semantic communication from the aspects of effectiveness and reliability.Finally,we present an example to illustrate the realization of semantic communication,which is proven effective in reducing data traffic compared to traditional communication mechanisms.

关 键 词：语义通信 智能体通信 语义相似度 语义编码 语义译码 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]