多项选择式机器阅读理解综述  

Review of multiple-choice machine reading comprehension

作　　者：马晨辉 施水才[1,2] 肖诗斌 MA Chenhui;SHI Shuicai;XIAO Shibin(Computer School,Beijing Information Science&Technology University,Beijing 100101,China;TRS Information Technology Co.Ltd,Beijing 100101,China)

机构地区：[1]北京信息科技大学计算机学院,北京100101 [2]拓尔思信息技术股份有限公司,北京100101

出　　处：《北京信息科技大学学报（自然科学版）》2021年第5期91-96,共6页Journal of Beijing Information Science and Technology University

摘　　要：对近几年迅速发展的多项选择式机器阅读理解的研究进展进行了归纳总结。首先介绍了多项选择式阅读理解的任务以及评价方法;其次介绍了当前多项选择式阅读理解的数据集;然后对多项选择式机器阅读理解的传统特征匹配方法进行了梳理,并概述了基于深度学习方法的模型架构和实现方法;最后对多项选择式机器阅读理解技术进行展望。The research progress of multiple-choice machine reading comprehension which has developed rapidly in recent years was summarized.Firstly,the tasks and evaluation methods of multiple-choice reading comprehension were described.Secondly,the current datasets of multiple-choice reading comprehension were introduced,then the methods of multiple-choice machine reading comprehension were sorted out,and the model architecture and implementation methods based on deep learning were summarized.Finally,the prospect of multi-choice machine reading comprehension technology was discussed.

关 键 词：机器阅读理解 自然语言处理 深度学习 预训练模型 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]