L220, 10:00 7.6.2011
In this talk I will introduce eye-tracking as a means to interact with computers, and in particular with problem-solving interfaces. We have conducted experimental investigations of the effects of gaze-based interaction on problem-solving strategies, performance, and user experience. Using gaze-augmented interaction changed the nature of the problem-solving strategies towards more internal planning activities, that in turn lead to better performance, fewer errors, more immersion, and increased user experience.
We applied machine learning methods to learn and predict the internal problem-solving strategies from the exhibited eye-movements. The results show that qualitatively different cognitive states can be inferred from overt visual attention patterns with accuracy highly over chance levels. Such system could be used for more intelligent and proactive interaction with future problem-solving interfaces.