Some key open questions in computational modeling of higher sensory cortex

Abstract: Propelled by advances in biologically inspired computer vision and artificial intelligence, the past five years have seen significant progress in using deep neural networks to model response patterns of neurons in visual cortex. In this talk, I'll first briefly review this progress in the context of my work, as well as those of others.  I'll then discuss what I think are a number of key ‘open questions’ that I believe will drive research in computational models of sensory systems over the next 5-10 years, both in visual cortex and beyond.   Ideally I'll be able to spark a lively discussion and draw everyone out on what they feel some key open questions are as well.