I received my Ph.D. in 1992 from the M.I.T. Dept. of Brain and Cognitive Sciences, and have been at Rutgers ever since. My main research interests are in visual perception, especially perceptual organization and shape; and in categorization and concept learning. In both these general areas my focus is on mathematical and computational models of human mental function.
In vision, I am interested in what makes human perceptual interpretations "make sense." Given a visual image, there are an infinite number of different ways to organize it, to group elements together, and to aggregate information in order to optimally estimate the structure of the world. The visual system is able to select from among these just that interpretation that seems most likely to provide a useful and accurate model of the world outside our heads. This idea has led me to Bayesian and other mathematical models that emphasize the maximization of simplicity in the organization of image data.
In categorization and concept learning, I am similarly interested in how the mind organizes groups of objects into coherent collections and hierarchies. In experimental work, we have found that human learners, given a set of objects to be learned, tend to form categories that are simple as possible. This idea opens up an enormous set of research questions about what perceptual features form the basis for categorization, how these features are selected in order to reduce representational complexity, and how these goals relate to the structure of the natural world.