CSE Distinguished Lecture: Daniel Lee

CSE Distinguished Lecture

Representation and Readout of Object Manifolds

The sixth speaker in the Fall 2016 CSE Distinguished Lecture Series is Daniel Lee, Director of the GRASP Laboratory and UPS Foundation Chair Professor in the School of Engineering and Applied Science at the University of Pennsylvania. Professor Lee will discuss the “Representation and Readout of Object Manifolds.”

Date: Monday, November 7
Time: 11:00am – Noon
Location: Room 1202, CSE Building

Host: CSE Prof. Lawrence Saul

Abstract: Objects are represented by sensory systems as continuous manifolds due to the sensitivity of neuronal responses to physical features such as location, orientation and intensity. Manifolds have also long been used in machine learning to model data variability due to pose, illumination, expression, and background changes. But how well can object manifolds be processed by neural networks? We generalize the traditional notions of linear separability and VC dimension from input points to input manifolds. A statistical mechanical analysis provides theoretical predictions for the capacity and margin of the optimal separating hyperplane for varying manifold dimensionalities, sizes and shapes. We also introduce a learning algorithm that estimates the parameters of the maximum margin classifier with guaranteed convergence that can be used to better understand how manifold representations are reformatted across the various layers of deep neural networks.

Bio: Daniel Lee is the UPS Foundation Chair Professor in the School of Engineering and Applied Science at the University of Pennsylvania. He received his B.A. summa cum laude in Physics from Harvard University and his Ph.D. in Condensed Matter Physics from the Massachusetts Institute of Technology in 1995. Before coming to Penn, he was a researcher at AT&T and Lucent Bell Laboratories in the Theoretical Physics and Biological Computation departments. He is a Fellow of the IEEE and AAAI and has received the National Science Foundation CAREER award and the University of Pennsylvania Lindback award for distinguished teaching. He was also a fellow of the Hebrew University Institute of Advanced Studies in Jerusalem, an affiliate of the Korea Advanced Institute of Science and Technology, and organized the US-Japan National Academy of Engineering symposium on Frontiers of Engineering. As director of the GRASP Laboratory and co-director of the CMU-Penn University Transportation Center, his group focuses on understanding general computational principles in biological systems, and on applying that knowledge to build autonomous systems.