Informing Computational Models?

[bethbailey]

Thank you for presenting!

Like many of my fellow students, I am interested in potential applications of your research to other avenues. Specifically, I wonder if you think your findings about rhythmic alpha frequency band activity could inform computational models that try to process visual scenes. More generally, how do you think specific findings about neural activity can/cannot or should/should not inform models for computer processing of complex things, such as moving scenes and images?

[edawh]

I think this approach could be a nice way to provide converging evidence for questions like: which aspects of a scene attract attention more readily, or for longer periods of time. we know that humans are quite efficient at understanding global scene properties, so learning about how they bias processing within scenes might provide useful hints for how to make computers better or more efficient at scene analysis.