计算机科学与探索 ›› 2022, Vol. 16 ›› Issue (2): 280-295.DOI: 10.3778/j.issn.1673-9418.2104042

• 综述·探索 • 上一篇    下一篇

面向端到端的情感对话生成研究综述

王春喻1, 马志强1,2,+(), 杜宝祥1, 贾文超1, 王洪彬1, 宝财吉拉呼1   

  1. 1.内蒙古工业大学 数据科学与应用学院,呼和浩特 010080
    2.内蒙古工业大学 内蒙古自治区基于大数据的软件服务工程技术研究中心,呼和浩特 010080
  • 收稿日期:2021-04-12 修回日期:2021-09-07 出版日期:2022-02-01 发布日期:2021-09-13
  • 通讯作者: + E-mail: mzq_bim@imut.edu.cn
  • 作者简介:王春喻(1997—),男,山西忻州人,硕士研究生,主要研究方向为自然语言处理、对话生成。
    马志强(1972—),男,内蒙古托克托人,硕士,教授,主要研究方向为多媒体信息处理、自然语言处理、机器学习。
    杜宝祥(1995—),男,山东济宁人,硕士研究生,主要研究方向为自然语言处理、对话生成。
    贾文超(1997—),男,河南周口人,硕士研究生,主要研究方向为自然语言处理、对话生成。
    王洪彬(1989—),男,山东菏泽人,硕士,讲师,主要研究方向为语音处理、自然语言处理、机器学习。
    宝财吉拉呼(1983—),男,内蒙古通辽人,博士,讲师,主要研究方向为机器学习、计算机视觉处理、生物信号处理、多媒体信息处理、自然语言处理。
  • 基金资助:
    国家自然科学基金(61762070);国家自然科学基金(61862048);内蒙古自然科学基金(2019MS06004);内蒙古自治区科技重大专项(2019ZD015);内蒙古自治区关键技术攻关计划项目(2019GG273);内蒙古自治区科技成果转化专项资金(2020CG0073)

Survey of Research on End-to-End Emotional Dialogue Generation

WANG Chunyu1, MA Zhiqiang1,2,+(), DU Baoxiang1, JIA Wenchao1, WANG Hongbin1, BAO Caijilahu1   

  1. 1. College of Data Science and Application, Inner Mongolia University of Technology, Hohhot 010080, China
    2. Inner Mongolia Autonomous Region Engineering & Technology Research Centre of Big Data Based Software Service, Inner Mongolia University of Technology, Hohhot 010080, China
  • Received:2021-04-12 Revised:2021-09-07 Online:2022-02-01 Published:2021-09-13
  • About author:WANG Chunyu, born in 1997, M.S. candidate. His research interests include natural language processing and dialogue generation.
    MA Zhiqiang, born in 1972, M.S., professor. His research interests include multimedia information processing, natural language processing and machine learning.
    DU Baoxiang, born in 1995, M.S. candidate. His research interests include natural language processing and dialogue generation.
    JIA Wenchao, born in 1997, M.S. candidate. His research interests include natural language processing and dialogue generation.
    WANG Hongbin, born in 1989, M.S., lecturer. His research interests include speech processing, natural language processing and machine learning.
    BAO Caijilahu, born in 1983, Ph.D., lecturer. His research interests include machine learning, computer vision processing, biological signal processing, multimedia information processing and natural language processing.
  • Supported by:
    National Natural Science Foundation of China(61762070);National Natural Science Foundation of China(61862048);Natural Science Foundation of Inner Mongolia(2019MS06004);Major Science and Technology Project of Inner Mongolia(2019ZD015);Key Technology Research Program of Inner Mongolia(2019GG273);Fund for Transformation of Scientific and Technological Achievements from Inner Mongolia(2020CG0073)

摘要:

人机对话作为人工智能的重要研究内容,受到了学术界和工业界的广泛关注。受到深度学习在自然语言处理成功应用的启发,越来越多的神经网络模型被研究者关注。其中基于端到端的神经网络模型能够从大规模语料中学习到有价值的规律和特征,生成有意义且多样性的回复,被广泛地应用于情感对话生成研究中。面向基于端到端模型的情感对话生成研究展开综述。首先,针对现有的研究成果,梳理并介绍了当前情感对话生成研究面向的任务和主要解决的问题,并且做出了详细的定义,整理并介绍了情感对话生成模型建模所需的数据集。其次,对端到端的神经网络模型的原理进行了简单的概述,并且分析和总结了情感对话生成研究在每个基础模型中的改进、研究现状、模型涉及的评价指标以及模型的性能。再次,对现阶段涉及到的模型评价方式按照自动评价以及人工评价方式进行了总结。最后,对未来情感对话生成研究的发展方向进行了展望。

关键词: 人机对话, 深度学习, 端到端, 情感对话生成

Abstract:

Human-machine dialogue, an important research component of artificial intelligence, has received widespread attention from academia and industry. Inspired by the successful application of deep learning in natural language processing, a growing number of neural network models are being focused on by researchers. Among them, end-to-end based neural network models are able to learn valuable patterns and features from large-scale corpus to generate meaningful and diverse responses, and are widely used in research on emotional dialogue generation. This paper presents a review of the research on end-to-end models for emotional dialogue generation. Firstly, the tasks and main problems addressed by current research on emotional dialogue generation are outlined and defined in detail in the light of existing research results. The datasets required for modeling emotional dialogue generation models are organized and presented. Secondly, a brief overview of the principles of end-to-end neural network models is given, and the improvements in each of the underlying models, the current state of research, the evaluation metrics involved in the models, and the performance of the models are analyzed and summarized. Thirdly, the evaluation methods involved in the current stage of model evaluation are summarized in terms of automatic evaluation as well as manual evaluation. Finally, this paper prospects the development direction of the research on the generation of emotional dialogue in the future.

Key words: human-machine dialogue, deep learning, end-to-end, emotional dialogue generation

中图分类号: