智能对话系统研究综述

发布时间：2018-04-15 14:10

本文选题：对话系统 + 深度学习　；参考：《北京工业大学学报》2017年09期

【摘要】：智能对话系统作为人工智能领域的核心技术,即将成为新的和谐人机交互方式,具有重大的研究意义和应用价值,因此,较为全面、深入地总结了深度学习模型在对话生成中的应用及对话系统领域的研究进展和现状.首先,阐述了智能对话系统的类型划分,介绍系统的模块框架构成,包括各模块的主要研究问题与关键技术的主流思路和研究现状;然后,从理论模型、研究进展及可用性等角度深度剖析了现有的对话生成技术解决方案,重点分析了应用于自然语言生成的序列到序列的神经网络结构及搭建原理;最后,对存在的问题进行总结,并展望了未来的研究方向.
[Abstract]:As the core technology in the field of artificial intelligence, intelligent dialogue system is about to become a new way of harmonious human-computer interaction, which has great research significance and application value, therefore, it is more comprehensive.The application of deep learning model in dialogue generation and the research progress and present situation of dialogue system are summarized.First of all, the classification of the intelligent dialogue system is described, and the structure of the system module framework is introduced, including the main research problems of each module and the main ideas and research status of the key technologies, and then, from the theoretical model,From the perspectives of research progress and usability, this paper deeply analyzes the existing dialog generation technology solutions, focusing on the analysis of the neural network structure and the construction principle of the sequence to sequence applied in natural language generation. Finally, the existing problems are summarized.The future research direction is also prospected.
【作者单位】：北京工业大学信息学部;联想集团北京研究院;
【基金】：国家自然科学基金资助项目(61370113) 北京市自然科学基金资助项目(4152005) 北京市教育委员会重点资助项目(KZ201610005009)
【分类号】：TP18;TP391.1

【相似文献】