摘要
问题生成是自然语言生成领域的子任务,该任务的目标是通过向机器提供一段文本,让机器生成与该文本相关的问题。问题生成有广泛的应用领域,如阅读理解、对话式问题生成、数据增强等,也因此引起了相关研究人员的关注和重视。本文首先对问题生成任务进行了简单的概述,包括其基本概念、研究意义,其次介绍了相关的数据集,进一步阐述了问题生成任务的技术发展路线,并对问题生成技术的应用方向进行了总结和分类,然后讨论了问题生成的一些前沿研究成果,最后介绍了问题生成的自动化评价指标,以及未来的研究方向和挑战。
Question generation is one of natural language generation tasks,which aims to generate questions by a given text.Question generation has a wide range of applications,such as reading comprehension,conversational question generation,data enhancement,etc.,which has attracted the attention of relevant researchers.This paper firstly gives a brief overview of the question generation task,including its basic concept and research significance,introduces relevant datasets and expounds the technical route of the question generation task,nextly summarizes and classifies the application direction of the question generation technology,and then introduces some frontier researches of question generation.Finally,the paper introduces the evaluation metric of question generation,the future research direction and challenge.
作者
祝胜泉
邵清
ZHU Shengquan;SHAO Qing(School of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)
出处
《智能计算机与应用》
2025年第3期170-180,共11页
Intelligent Computer and Applications
关键词
神经网络
问题生成
自然语言生成
机器阅读理解
评价指标
neural network
question generation
natural language generation
machine reading comprehension
evaluation metric
