![]() |
Jonathan Shaw (Graduate Student ) |
UR > CS > Research > Vision > People > Jonathan Shaw |
![]() | Contact:Jonathan ShawCSB 630 Department of Computer Science University of Rochester Rochester, NY 14627 vox: (585) 275-2569 email: jshaw@cs.rochester.edu home: http://www.cs.rochester.edu/jshaw vision: http://www.rochester.edu/research/vision/people/Jonathan_M_Shaw |
About:Jonathan is in his second year of the Computer Science Ph. D. program at the University of Rochester. He is interested in machine learning, and focussing on automatic finding of temporal invariance. His other interests include theology, his fiancee, and piano. | |
Research Interests:I am chiefly interested in understanding and modelling hwo te brain works, which leads me to Machine Learning and Neural Models of Computation. As the operation of the brain is largely unsupervised, I focus in unsupervised learning. Specifically, I have a growing interest in developing unsupervised methods of finding temporal invariance. I am currently exploring the possibility that the brain is targeted at finding properties of its input which change rarely, as opposed to slowly. In my present work, my models presuppose that the most interesting things in the world are discreet. For example, at any given instant, I am interested in whether I am talking to my advisor or my sister. This is a discreet property, as I am never talking to a mixture of them, although I may conceivably be talking to both. This hypothesis allows for the development of algorithms which provide descriptions of the world based on the probabilities that individual features of the world which rarely change are present at a given point in time. | |
Projects: | |
Publications: |