Temporal and spatial design of explanations in a multimodal system

Abstract

Modern dialog systems are known to act user-specific. They apply individual decisions for content presentation and course adaptation. However, it is still an open research question how additional, but required explanations should be integrated best into a given dialog structure. Previous research focused on the improvement of user knowledge models and its fine-grained use in human-computer interaction, but does not directly address the temporal and spatial aspects of presentation when it comes to explanations. In this paper, we introduce different strategies for an ad-hoc integration of required explanations. We describe a user study, and show which parameters from the fields of user experience, personality, and cognitive load theory have what effects on the applied strategies. We expect that our findings can help to increase usability and decrease unwanted cognitive load.