Toward a Task-driven Intelligent GUI Adaptation by Mixed-initiative

Adapting the user interface (UI) to the changing context of use is intended to support the interaction effectiveness and sustain UI usability. However, designing and/or processing UIs adaptation at design time does not encompass real situation requirements. Adaptation should have a cross-cutting and low-cost impact on software patterning and appearance with regard to the situation and the ambient-context. To meet this requirement, we present TADAP proposal for run-time adaptive and adaptable UI based user feedbacks and machine learning. It allows a task-driven adaptation of the user interface (UI) at runtime by mixed-initiative. The particularity of TADAP is the utilization of Machine Learning potential to support context-aware runtime adaptation within model-driven UI. Further, TADAP allows the UI adaptation by mixed-initiative (User and System) considering the user preferences (implicit and explicit) during an interaction. Such a mixed-initiative runtime UI-adaptation tool provides recommendations on how to personalize the UI. Further, it has the ability to track real-time users' interventions and learn their preferences. Diverse tests were performed and showed TADAP as a promising initiative for intelligent model-driven UI adaptation.

Automated summary

This paper discusses the proposal of TADAP for achieving runtime adaptive UI through user feedback and machine learning, which allows for cross-cutting impact of adaptation, taking into account user preferences and real-time context.