Categorical Perception: A Groundwork for Deep Learning

Classification is one of the major tasks that deep learning is successfully tackling. Categorization is also a fundamental cognitive ability. A well-known perceptual consequence of categorization in humans and other animals, categorical per ception, is notably characterized by a within-category compression and a between-category separation: two items, close in input space, are perceived closer if they belong to the same category than if they belong to different categories. Elaborating on experimental and theoretical results in cognitive science, here we study categorical effects in artificial neural networks. We combine a theoretical analysis that makes use of mutual and Fisher information quantities and a series of numerical simulations on networks of increasing complexity. These formal and numerical analyses provide insights into the geometry of the neural representation in deep layers, with expansion of space near category boundaries and contraction far from category boundaries. We investigate categorical representation by using two complementary approaches: one mimics experiments in psychophysics and cognitive neuroscience by means of morphed continua between stimuli of different categories, while the other introduces a categoricality index that, for each layer in the network, quantifies the separability of the categories at the neural population level. We show on both shallow and deep neural networks that category learning automatically induces categorical perception. We further show that the deeper a layer, the stronger the categorical effects. As an outcome of our study, we propose a coherent view of the efficacy of different heuristic practices of the dropout regularization technique. More generally, our view, which finds echoes in the neuroscience literature, insists on the differential impact of noise in any given layer depending on the geometry of the neural representation that is being learned, that is, on how this geometry reflects the structure of the categories.

Autism and schizophrenia: Similar perceptual consequence, different neurobiological etiology?

Phillips & Silverstein (P&S, 2003) propose that NMDA-receptor dysfunction may be the fundamental neurobiological mechanism underlying and associating impaired holistic perception and cognitive coordination with schizophrenic psychopathology. We discuss how the P&S hypothesis shares different aspects of the weak central coherence account of autism from both theoretical and experimental perspectives. Specifically, we believe that neither those persons with autism nor those with schizophrenia integrate visuo-perceptual information efficiently, resulting in incongruous internal representations of their external world. However, although NMDA-hypofunction may be responsible for perceptual impairments in schizophrenia and possibly autism, we suggest that it is highly unlikely that NMDA-hypofunction is specifically responsible for the autistic behavioral symptomology, as described by P&S in their target article.