汉语口语对话系统中语义分析的消歧策略  

The Disambiguation Strategy of Semantic Analysis in Spoken Dialogue Systems

作　　者：刘蓓[1] 杜利民[1] 

机构地区：[1]中国科学院声学研究所语音交互信息技术研究中心,北京100080

出　　处：《中文信息学报》2005年第1期76-83,共8页Journal of Chinese Information Processing

基　　金：国家 973重点基础研究发展项目资助 (G19980 30 5 0 5 )

摘　　要：框架语义分析是目前汉语口语对话系统中常用的语义解析方法 ,本文分析了语义分析过程中容易产生的两种典型歧义现象 -结构歧义和语义关系歧义。并针对这两种歧义结构 ,分别提出基于语义PCFG模型的结构歧义消歧策略以及基于语义期待模型EM的语义关系歧义消歧策略 ,并给出了有效的消歧算法。实验结果表明综合运用本文提出的消歧策略后 ,基线系统理解模块的句子语义分析正确率大大提高 ,从原来的75 7%上升到 91 5 % ,而且标志语义单元理解率的三项指标 ,准确率 ,召回率和精度也平均提高了 10 %。Frame semantic analysis is one of the most commonly used semantic analysis method in Chinese spoken dialogue systems research. And the two typical ambiguous structures encountered in semantic analysis are outer-ambiguity and structural-ambiguity. According to the features of these two ambiguous structures, this paper puts forth the semantic PCFG model based disambiguation strategy to solve structural-ambiguity and the semantic Expectation Model (EM) integrated disambiguation strategy to solve outer-ambiguity. Efficient algorithms of these two methods are also provided. The experimental results show that synthetically use of these two disambiguation methods can most greatly improve the performance of the understanding module of the base-line system. The sentence accuracy is improved from 75.7% to 91.5%, and the three targets of semantic understanding rate-correction, recall, and precision are also improved 10% averagely.

关 键 词：计算机应用 中文信息处理 口语对话系统 语义分析 消歧 算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]