Exploring Visual Representation in a Task using VAEs

Our group has developed hierarchical VAE models that can capture top-down effects of biological vision. We aim to extend our model to predict neuronal responses during a behavioral tasks when an animal performs a discrimination task. Neuronal responses to a given visual stimulus in the cortex depend on contextual factors. We will develop a method that enables computationally efficient determination of the posterior in a VAE framework in response to a dynamically changing prior.