北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2022, Vol. 45 ›› Issue (6): 12-21.

• 论文 • 上一篇    下一篇

面向智能机器通信的语义信息刻画及度量

马楠1,宋孟书2,刘宜明3,董辰2   

  1. 1. 北京邮电大学无线新技术研究所
    2. 北京邮电大学
    3. 北京邮电大学信息与通信工程学院
  • 收稿日期:2022-06-22 修回日期:2022-09-15 出版日期:2022-12-28 发布日期:2022-11-24
  • 通讯作者: 宋孟书 E-mail:MSSong@bupt.edu.cn

Description and measurement of semantic information for the intelligent machine communication

  • Received:2022-06-22 Revised:2022-09-15 Online:2022-12-28 Published:2022-11-24
  • Contact: Mengshu Song E-mail:MSSong@bupt.edu.cn

摘要: 语义信息具备高度抽象、智能简约等特性,语义通信引入语义新维度,关注信息内容而非编码符号,将有效提高信息的传输效率。随着人工智能与通信技术的深度融合,语义通信得到学术界和工业界广泛关注,将有助于解决智能机器通信中超大规模连接、巨量数据传输等难题,对实现机器间高效通信发挥重要作用。以信息论为基础、人工智能方法为指导,概述端到端语义通信系统设计方法,总结分析了语义信息的刻画及度量、语义信息编解码方法和语义通信系统评价指标。

关键词: 语义信息, 语义通信, 智能通信, 智能机器

Abstract: Semantic information has the characteristics of high abstraction, intelligence and simplicity. Semantic communication introduces a new semantic dimension and focuses on information content rather than coding symbols, which will greatly improve the efficiency of information transmission. With the deep integration of artificial intelligence and communication technology, semantic communication has been widely concerned by the academic and industrial circles, which will help solve the problems of super-large-scale connection and massive data transmission in intelligent machine communication, and play an important role in realizing efficient communication between machines. Based on information theory and guided by artificial intelligence, this paper summarizes the design method of end-to-end intelligent semantic communication system, presents the description and measurement of semantic information, investigates the encoding and decoding methods and analyses the evaluation indexes of semantic communication system.

Key words: Semantic information, Semantic communication, Intelligent communication, Intelligent Machine

中图分类号: