Mathematics Colloquia and Seminars

Return to Colloquia & Seminar listing

Our percepts of the world are clearly *inferred*, rather than being computed directly from the available data. This means that our brains must be endowed with powerful inferential machinery - i.e., probabilistic models - for combining incoming sensory information together with prior knowledge in order to infer what's "out there" in the environment. In this talk I will present a simple version of a probabilistic model for primary visual cortex (V1) that is based on the idea of sparse coding - i.e., where images are represented by a small number of active units at any given time. I will then present the results of computational simulations showing that this idea is consistent with the receptive field properties found in V1 neurons, and I will present data supporting the idea that cortical neurons are attempting to infer sparse representations of images. Both the model and the data make clear that if we are to actually understand what is going on the cortex, we need to focus our efforts on studying how it operates under natural conditions.