Learning to combine foveal glimpses with a third-order Boltzmann machine
Published on Jan 12, 20114769 Views
We describe a model based on a Boltzmann machine with third-order connections that can learn how to accumulate information about a shape over several fixations. The model uses a retina that only has e