Emotion Recognition Based Preference Modelling in Argumentative Dialogue Systems

Abstract

Within this work, we present an approach to model the opinion of a human towards a specific topic in a fine-grained way by using weighted bipolar argumentation graphs. In addition, we discuss how the therefore required rating of related aspects can be collected by means of emotion recognition techniques and discuss an application scenario based on the state-of-the-art Argumentative Dialogue System EVA in which the proposed techniques can be applied.