Towards a methodology for the evaluation of Multimodal Interactive Systems: case studies
Evaluating the usability and the quality of Multimodal Interactive Systems is a complex endeavour. Many factors have to be taken into account, and these factors depend strongly on the nature of the application. The main task, therefore, is to determine a specific taxonomy of quality aspects for each particular Multimodal Interactive System. The task would be made much more tractable if it were possible to develop a general methodology - covering evaluation procedures, annotation (standards, tools..) and techniques to handle and interpret the information obtained – with which to derive the appropriate taxonomies for each context.
Towards these goals, our contribution has been directed to specific cases of study, and to obtain insight on what the best procedures and practices might be helpful. In particular, I will present our work in the field of biometrics, in the educational software for children with Special Needs, and the study that we plan to undertake on Spoken Language Dialogue Systems guided by an Embodied Conversational Agent.