移动通信2025，Vol.49Issue(7)：31-35,60,6.DOI:10.3969/j.issn.1006-1010.20250524-0002
语义信息论视角下的知识库表征极限与压缩原理
Knowledge Base Representation and Compression Limits From a Semantic Information-Theoretic Perspective
摘要
Abstract
This paper investigates the fundamental limits of knowledge base representation and compression from the perspective of semantic information theory.A semantic rate-distortion function is introduced to quantify the tradeoff between compression efficiency and representational fidelity.Building on synonymic mapping relations in semantic information theory,we consider both fully aligned and mismatched semantic knowledge bases at the transmitter and receiver.A semantic distortion function and a knowledge base alignment metric are further defined to model and evaluate the impact of semantic consistency.Simulation results on the CUB dataset demonstrate that larger knowledge bases enable more efficient semantic compression.Moreover,a higher degree of alignment between transmitter and receiver knowledge bases significantly enhances semantic communication performance,thereby validating the effectiveness of the proposed theoretical framework and algorithmic design.关键词
语义率失真函数/语义知识库/语义信息论/压缩极限Key words
semantic rate-distortion function/semantic knowledge base/semantic information theory/compression limits分类
信息技术与安全科学引用本文复制引用
韩雨欣,牛凯,孙亚萍,刘洋,马楠,崔曙光..语义信息论视角下的知识库表征极限与压缩原理[J].移动通信,2025,49(7):31-35,60,6.基金项目
国家自然科学基金"语义信息的表征与传输理论"(62293481) （62293481）
国家自然科学基金"面向未来无线通信的高性能极化编码理论与方法"(62471054) （62471054）
国家自然科学基金"语义知识库驱动的零样本多层级语义编码与特征传输研究"(62301471) （62301471）
国家自然科学基金"语义知识库构建方法与智能演进机理"(62293482) （62293482）
移动信息网络国家科技重大专项(2024ZD1300700) （2024ZD1300700）
