Laκsa mobile app for intelligible context-aware interruption management.
Architecture of the Intelligibility Toolkit for query processing, automatic explanation generation, and visualization.
Validation applications developed with the Intelligibility Toolkit.

Sensors and artificial intelligence can allow apps to be pro-active and adaptive. These can provide personalized and context-aware services, such as automatically silencing our phones depending on where we are and what we are doing, or automatically detecting anomalous events regarding elderly loved ones. As promising as these services and apps are, there remains an issue of how to understand them when they misbehave. We research on how to improve understanding and trust in context-aware applications by eliciting requirements for explanations, developing techniques to automatically generate explanations, and evaluating the positive impact and compromises of explanations.

Relevant papers: