面向移动终端的人机对话系统

发布时间：2018-01-06 15:40

本文关键词：面向移动终端的人机对话系统　出处：《桂林电子科技大学》2017年硕士论文　论文类型：学位论文

【摘要】：人机对话技术是人机交互领域重要交互手段之一,在当今智能化生活和产业领域具有广泛应用,例如智能家居、陪护机器人、智能客服、手机助手等。本文对人机对话中用户教授意图下的答案反馈方法进行研究,旨在探索一种教授意图下交流体验度更佳的答案反馈方法。首先,本文介绍了人机对话系统的研究意义及国内外研究现状;然后,对当前比较流行的对话管理方法和人机对话相关技术进行了综述性介绍;最后,以用户教授意图下的答案反馈作为研究对象,提出了一种基于自我对话的面向用户教授意图答案反馈方法,并实现了一个面向移动终端的人机对话系统。本文的工作主要有以下2个方面:(1)提出了一种基于自我对话的面向用户教授意图的答案反馈方法。传统的对话系统在用户教授意图下,通常返回模棱两可的答案,不利于对话的进一步进行。针对此问题,在明确用户教授意图情况下,首先根据对话历史选出当前对话进程对应的问题,根据匹配到的问题,结合用户教授答案及互联网相关信息模拟多轮自我对话;然后对自我对话获取的多轮信息进行过滤和摘要抽取;最终,将抽取得到的摘要作为系统对用户教授意图的反馈。实验结果表明该方法能够在用户教授意图下学习用户对话中的知识,并在一定程度上提高了用户对话的体验度。(2)采用WebService技术,基于Android系统平台实现了一个面向移动终端的人机对话系统。该对话系统能够记录多轮对话历史信息并根据对话历史进行多轮连续对话;系统支持天气、餐饮、景点、路线查询,同时也支持开放式的自由聊天,能够在与用户对话过程中从网络和用户的对话中获取新的知识;同时,系统基于WebService技术实现,具有很强的移植性和跨平台能力。
[Abstract]:The technology of man-machine dialogue is one of the important interactive means in the field of human-computer interaction. It has been widely used in the field of intelligent life and industry, such as smart home, escort robot, intelligent customer service. Mobile phone assistant and so on. This paper studies the answer feedback method of user's intention in man-machine dialogue, in order to explore a kind of answer feedback method with better communication experience under the professor's intention. First of all. This paper introduces the research significance of man-machine dialogue system and the current research situation at home and abroad. Then, the current popular dialogue management methods and man-machine dialogue related technologies are summarized. Finally, taking the answer feedback under the intention of the user professor as the research object, this paper proposes a method based on self-dialogue to feedback the answer to the user professor intention. A man-machine dialogue system for mobile terminal is implemented. The main work of this paper is as follows: 1). This paper presents an answer feedback method based on self-dialogue, which is oriented to the intention of the user professor. The traditional dialogue system is based on the intention of the user professor. The return of ambiguous answers is usually not conducive to the further development of the dialogue. In view of this problem, the current dialogue process is first selected according to the history of the dialogue in the case of the user's intention to teach. According to the matching questions, combined with the user to teach the answers and Internet related information simulation of multiple rounds of self-dialogue; Then the multi-round information obtained by self-dialogue is filtered and abstracted. Finally, the abstract is taken as the feedback of the user's intention. The experimental results show that the method can learn the knowledge of the user's dialogue under the user's intention. And to a certain extent, improve the user dialogue experience degree. 2) the use of WebService technology. Based on the Android platform, a man-machine dialogue system for mobile terminal is implemented. The dialogue system can record the historical information of multi-round dialogue and carry out multi-round continuous dialogue according to the history of dialogue. The system supports weather, catering, scenic spots, route inquiry, and open free chat, which can gain new knowledge from the network and the user's dialogue. At the same time, the system is based on WebService technology and has strong portability and cross-platform ability.
【学位授予单位】：桂林电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP11

【参考文献】