tailieunhanh - Báo cáo khoa học: "Co-training for Predicting Emotions with Spoken Dialogue Data"

Natural Language Processing applications often require large amounts of annotated training data, which are expensive to obtain. In this paper we investigate the applicability of Co-training to train classifiers that predict emotions in spoken dialogues. In order to do so, we have first applied the wrapper approach with Forward Selection and Naïve Bayes, to reduce the dimensionality of our feature set. Our results show that Co-training can be highly effective when a good set of features are chosen. . | Co-training for Predicting Emotions with Spoken Dialogue Data Beatriz Maeireizo and Diane Litman and Rebecca Hwa Department of Computer Science University of Pittsburgh Pittsburgh Pa 15260 . beamt@ litman@ hwa@ Abstract Natural Language Processing applications often require large amounts of annotated training data which are expensive to obtain. In this paper we investigate the applicability of Co-training to train classifiers that predict emotions in spoken dialogues. In order to do so we have first applied the wrapper approach with Forward Selection and Naive Bayes to reduce the dimensionality of our feature set. Our results show that Co-training can be highly effective when a good set of features are chosen. 1 Introduction In this paper we investigate the automatic labeling of spoken dialogue data in order to train a classifier that predicts students emotional states in a human-human speech-based tutoring corpus. Supervised training of classifiers requires annotated data which demands costly efforts from human annotators. One approach to minimize this effort is to use Co-training Blum and Mitchell 1998 a semi-supervised algorithm in which two learners are iteratively combining their outputs to increase the training set used to re-train each other and generate more labeled data automatically. The main focus of this paper is to explore how Cotraining can be applied to annotate spoken dialogues. A major challenge to address is in reducing the dimensionality of the many features available to the learners. The motivation for our research arises from the need to annotate a human-human speech corpus for the ITSPOKE Intelligent Tutoring SPOKEn dialogue System project Litman and Silliman 2004 . Ongoing research in ITSPOKE aims to recognize emotional states of students in order to build a spoken dialogue tutoring system that automatically predicts and adapts to the student s emotions. ITSPOKE uses supervised learning to predict .