Đang chuẩn bị nút TẢI XUỐNG, xin hãy chờ
Tải xuống
Speaker’s intention prediction modules can be widely used as a pre-processor for reducing the search space of an automatic speech recognizer. They also can be used as a preprocessor for generating a proper sentence in a dialogue system. We propose a statistical model to predict speakers’ intentions by using multi-level features. Using the multi-level features (morpheme-level features, discourselevel features, and domain knowledge-level features), the proposed model predicts speakers’ intentions that may be implicated in next utterances. . | Speakers Intention Prediction Using Statistics of Multi-level Features in a Schedule Management Domain Donghyun Kim Diquest Research Center Diquest Inc. Seoul Korea kdh2007@sogang.ac.kr Hyunjung Lee Computer Science Engineering Sogang University Seoul Korea juvenile@sogang.ac.kr Choong-Nyoung Seon Computer Science Engineering Sogang University Seoul Korea wilowisp@gmail.com Harksoo Kim Computer Communications Engineering Kangwon National University Chuncheon Korea nlpdrkim@kangwon.ac.kr Abstract Speaker s intention prediction modules can be widely used as a pre-processor for reducing the search space of an automatic speech recognizer. They also can be used as a preprocessor for generating a proper sentence in a dialogue system. We propose a statistical model to predict speakers intentions by using multi-level features. Using the multi-level features morpheme-level features discourselevel features and domain knowledge-level features the proposed model predicts speakers intentions that may be implicated in next utterances. In the experiments the proposed model showed better performances about 29 higher accuracies than the previous model. Based on the experiments we found that the proposed multi-level features are very effective in speaker s intention prediction. 1 Introduction A dialogue system is a program in which a user and system communicate in natural language. To understand user s utterance the dialogue system should identify his her intention. To respond his her question the dialogue system should generate the counterpart of his her intention by referring to dialogue history and domain knowledge. Most previous researches on speakers intentions have been focused on intention identification techniques. On the contrary intention prediction techniques have been not studied enough although Jungyun Seo Computer Science Engineering Sogang University Seoul Korea seoj y@sogang.ac.kr there are many practical needs as shown in Figure 1. Example 1 Prediction of user s .
