A novel approach for data fusion and dialog management in user-adapted multimodal dialog systems

Multimodal dialog systems have demonstrated a high potential for more flexible, usable and natural humancomputer interaction. These improvements are highly dependent on the fusion and dialog management processes, which respectively integrates and interprets multimedia multimodal information and decides the next system response for the current dialog state. In this paper we propose to carry out the multimodal fusion and dialog management processes at the dialog level in a single step. To do this, we describe an approach based on a statistical model that takes user´s intention into account, generates a single representation obtained from the different input modalities and their confidence scores, and selects the next system action based on this representation. The paper also describes the practical application of the proposed approach to develop a multimodal dialog system providing travel and tourist information.