Section: New Results

Adaptive interfaces

Participant : Sylvain Malacria.

As news is increasingly accessed on smartphones and tablets, the need for personalizing news applications is apparent. In [19] , [20] , we report on a series of studies addressing key issues in the development of adaptive news interfaces. We first surveyed users’ news reading preferences and behaviors. We then implemented and deployed an Android application that logs users’ interactions with the application. We used the logs to train a classifier and showed that it is able to reliably recognize a user according to their reader type. Finally we evaluated alternative, adaptive user interfaces for each reader type. The evaluation demonstrates the differential benefit of the adaptation for different users and the feasibility of adaptive interfaces for news applications.

In [12] , we investigate the use of a companion application on a tablet to augment viewing of information-rich television programs. The application displays a synchronized graphical abstraction of the program’s content in the form of a concept map. Two experiments were conducted involving participants watching an astronomy documentary. Results show that the companion application improved participants’ understanding and recall of the program. Participants were found to manage their visual attention systematically when using the companion application, and correlations were found in the way they shifted their gaze from TV screen to tablet and back in response to changes in the program content. Increasing interaction with the application disrupted understanding of the television program and visual attention. Participants were positive about the value of companion applications for understanding and recall of programs, but distraction and “knowing where to look” were significant concerns.