基于多策略混合问答系统模型的图书馆咨询机器人的设计与应用  

Design and Application of Library Consultation Robot Based on Multi-strategy Mixed Question Answering System Model

作　　者：刘泽 徐潇洁 邵波[1,2] Liu Ze;Xu Xiaojie;Shao Bo

机构地区：[1]南京大学信息管理学院,江苏南京210023 [2]南京大学图书馆,江苏南京210023

出　　处：《新世纪图书馆》2022年第5期43-49,共7页New Century Library

摘　　要：随着人工智能技术的发展,智能咨询机器人可作为智慧图书馆咨询服务的新选择。论文从图书馆咨询服务的实际需求出发,搭建图书馆智能咨询机器人系统框架,基于知识库、文本硬匹配、相似度匹配、深度学习等技术,分别构建面向寒暄类和咨询类的问答子系统和面向任务类的对话子系统,并研究其应用效果。图书馆通过多策略混合的问答系统模型构建智能咨询机器人,能够提高问答系统的性能,为读者提供更优质准确的参考咨询服务。With the development of artificial intelligence technology,intelligent consulting robots can be used as a new choice for library consulting services.The framework of intelligent library consulting robots was built according to the actual needs of library consulting services.Based on repository text hard matching,similarity matching and deep learning,a question-and-answer subsystem for greetings and consulting and a dialogue subsystem for task-oriented dialogues were constructed respectively.This paper analyzed the effect of its application.An consulting robot platform is built through a multi-strategy mixed question answering system model,which can improve the performance of the question answering system and provide readers with more high-quality and accurate reference consulting services.

关 键 词：多策略混合 咨询服务 智能咨询机器人 系统框架 

分 类 号：G251[文化科学—图书馆学]