Influence of the data codification when applying evolving classifiers to develop spoken dialog systems

Abstract

In this paper we present a study of the influence of the representation of the data when applying evolving classifiers in a specific classification task. In particular, we consider an evolving classifier for the development of a spoken dialog system interacting in a practical domain. In order to conduct this study, we will first introduce an approach based on evolving fuzzy systems (EFS) which is employed to select the next system action of the dialog system. This classifier takes into account a set of evolving fuzzy rules which are automatically obtained using evolving systems. The reason for using EFS in this domain is that we can process streaming data on-line in real time and the structure and operation of the dialog model can dynamically change by considering the interaction of the dialog system with its users. Since we want to apply this evolving approach in a real domain, our proposal considers the data supplied by the user throughout the complete dialog history and the confidence measures provided by the recognition and understanding modules of the system. The paper is focused on the study of the influence of the codification of this input data to achieve the best performance of the proposed approach. To do this, we have completed this study for a real spoken dialog system providing railway information.

A Neural Network Approach to Intention Modeling for User-Adapted Conversational Agents

“…In our previous work we have evaluated different classifiers widely used in the field of natural language processing and speech technologies [ 6 – 8 ]. Neural networks have outperformed the other classifiers in different application domains [ 6 , 45 – 48 ].…”

“…The current paper presents the following important contributions. One of the main ideas of our proposal is that the dialogue system and the user model interact simultaneously, not offline as in our previous works [ 7 , 8 ], so that the user model is employed in real time by the dialogue manager to decide the best answer dynamically and not in a posteriori laboratory evaluations. In addition, the technology used to build the proposed user model does not replicate that of the dialogue manager.…”