摘要
口语理解是对话系统重要的功能模块,语义槽填充和意图识别是面向任务口语理解的两个关键子任务。近年来,联合识别方法已经成为解决口语理解中语义槽填充和意图识别任务的主流方法,介绍两个任务由独立建模到联合建模的方法,重点介绍基于深度神经网络的语义槽填充和意图识别联合建模方法,并总结了目前存在的问题以及未来的发展趋势。
Spoken Language Understanding(SLU), which includes two key sub-tasks slot filling and intent detection, is an important function module of the dialogue system. In recent years, joint recognition methods to solve slot filling and intent detection have become the mainstream methods of SLU. This paper introduces the methods of two sub-tasks, which develop from independent modeling to joint modeling. It focuses on the joint modeling methods based on deep neural network, analyzes current problems and future development trend of two sub-tasks.
作者
侯丽仙
李艳玲
李成城
HOU Lixian;LI Yanling;LI Chengcheng(College of Computer Science and Technology, Inner Mongolia Normal University, Hohhot 010022, China)
出处
《计算机工程与应用》
CSCD
北大核心
2019年第11期7-15,共9页
Computer Engineering and Applications
基金
国家自然科学基金(No.61562068,No.11704229,No.61640204,No.61806103)
内蒙古自然科学基金(No.2017MS0607)
内蒙古民委蒙古文信息化专项扶持子项目(No.MW-2014-MGYWXXH-01)
内蒙古自治区“草原英才”工程青年创新创业人才项目
内蒙古师范大学研究生创新基金(No.CXJJS18112)
