• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>人工知能学会志 >Affect Analysis of Textual Input Utterance in Japanese and Its Application in Human-Computer Interaction
【24h】

Affect Analysis of Textual Input Utterance in Japanese and Its Application in Human-Computer Interaction

机译：日语文本输入话语的情感分析及其在人机交互中的应用

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

In this dissertation I present my research on enhancing machines with Emotional Intelligence. I develop a set of affect analysis tools and propose methods for efficient utilization of emotive information obtained by these tools. The first system developed is ML-Ask, a system for affect analysis of textual input utterance in Japanese. ML-Ask first separates emotive utterances from non-emotive and in the emotive utterances seeks for expressions of specific emotion types. The second system, CAO, is a system for analysis of emoticons. CAO extracts emoticons from input and determines the specific emotion types they express. Firstly, it matches the extracted emoticons to a predetermined emoticon database containing over ten thousand emoticon samples extracted from the Web and annotated automatically. The emoticons difficult to deal with this way are automatically divided into semantic areas representing mouths" or "eyes" and matched to a semantic area database created and annotated automatically. The above systems are then utilized in two methods for enhancing of Human-Computer Interaction. The first is a method for automatic evaluation of conversational agents. In this method, the affect analysis systems are used to analyze users' emotional engagement during conversation. This data is reinterpreted to specify general attitudes to the conversational agent and its performance. The method revealed similar tendencies to a standard usability questionnaire, proving its applicability in automatic evaluation of Japanese-speaking conversational agents. Finally, I present a method for determining whether emotions expressed by speaker are appropriate for the context of the conversation. In this method, affect analysis systems estimate the speaker's affective states and a Web mining technique gathers from the Internet emotive associations consisting of a list of emotions that should be expressed at the moment. Implementing the above methods to a conversational agent could allow it choose appropriate conversational procedures and therefore enhance human-computer interaction. I conclude the dissertation with a discussion on possible applications for the proposed systems and methods, and further work needed for implementation of the full scope of Emotional Intelligence in machines.
机译：在这篇论文中，我介绍了我对用情商增强机器的研究。我开发了一套情感分析工具，并提出了有效利用这些工具获得的情感信息的方法。第一个开发的系统是ML-Ask，这是一种用于日语文本输入发音的情感分析的系统。 ML-Ask首先将情绪发声与非情绪发声分开，然后在情绪发声中寻求特定情绪类型的表达。第二个系统，CAO，是一个用于分析图释的系统。 CAO从输入中提取表情符号，并确定它们表达的特定情感类型。首先，它将提取的图释与预定的图释数据库进行匹配，该数据库包含从Web提取并自动注释的一万多个图释样本。将难以处理的表情符号自动划分为代表“嘴巴”或“眼睛”的语义区域，并与自动创建和注释的语义区域数据库进行匹配，然后将上述系统用于增强人机交互的两种方法中。第一种是自动评估对话主体的方法，该方法使用情感分析系统来分析用户在对话过程中的情感参与度，并重新解释此数据以指定对对话主体及其性能的一般态度。类似于标准可用性调查表的趋势，证明了其适用于日语会话代理人的自动评估;最后，我提出了一种确定说话者表达的情绪是否适合会话环境的方法，该方法可以影响分析系统估计说话者的情感状态和网络挖掘技术从互联网上的情绪关联中收集信息，其中包括目前应表达的一系列情感。将以上方法应用于会话代理可以使其选择适当的会话过程，从而增强人机交互性。最后，本文讨论了所提出的系统和方法的可能应用，以及在机器中实现完整范围的情商所需的进一步工作。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号