面向查询的文本摘要关键技术研究综述  

Research Overview of Key Techniques for Query-Focused Summarization

作　　者：徐睿 刘金[1] 亚静 XU Rui;LIU Jin;YA Jing(National Computer System Engineering Research Institute of China,Beijing 102209,China;Northking In f ormation Technology Co.,Ltd.,Beijing 100081,China;School of Intelligence Science and Technology,Peking University,Beijing 100871,China)

机构地区：[1]华北计算机系统工程研究所,北京102209 [2]京北方信息技术股份有限公司,北京100081 [3]北京大学智能学院,北京100871

出　　处：《软件工程》2024年第2期1-9,共9页Software Engineering

摘　　要：面向查询的文本摘要是自动文摘中的一个特殊领域,可以根据用户个性化的查询需求,从原始文档或文档集中提取有价值的摘要信息。目前,该技术已经在面向查询的搜索引擎、智能化信息检索、问答系统等领域得到广泛应用,并受到越来越多的关注。文章基于面向查询的文本摘要任务的典型技术框架,从查询理解、文档处理和信息组织三个方面对其国内外研究方法的现状进行对比和分析,对不同业务场景的应用进行了举例,归纳了面向查询的文本摘要面临的挑战及发展趋势。Query-Focused Summarization(QFS)is a special area of automatic summarization,which can be used to extract valuable summary information from original document or multi-documents based on the user's personalized query requirements.The technology has been widely used in the search engine,intelligent information retrieval,Q&A system and other areas,and has received more and more attention.In this paper,we start from the typical technical framework for achieving this task,compare and analyze the existing research methods in three aspects:query comprehension,document processing and information organization,and summarize the challenges and development trends QFS.

关 键 词：面向查询 文本摘要 自然语言处理 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]