Visit ShanghaiTech University | 中文 | How to find us
HOME > News and Events > Events
Structured Sparse Subspace Clustering and Some Extensions
Date: 2017/11/27             Browse: 52

Speaker:     Associate Prof. Chun-Guang Li, BUPT

Time:          Nov  27,    14:00 – 15:00

Location:    Room 1A-200, SIST Building

Host:          Prof. Manolis Tsakiris


High-dimensional data often reside in low-dimensional structure, e.g. subspaces.  Segmenting data drawn from a union of subspaces refers to the problem of subspace clustering. State-of-the-art approaches for solving this problem follow a two-stage approach. In the first step, an affinity matrix is learned from the data. In the second step, the segmentation is found by applying spectral clustering to this affinity. While this approach has led to state-of-the-art results in many applications, it is sub-optimal because it does not exploit the fact that the affinity and the segmentation depend on each other. This talk will present a joint optimization framework --- Structured Sparse Subspace Clustering --- for learning both the affinity and the segmentation. The framework is based on expressing each data point as a structured sparse linear combination of all other data points, where the structure is induced by a norm that depends on the unknown segmentation. Moreover, the framework is extended into the setting of constrained subspace clustering --- in which partial side-information is available, and the setting of semi-supervised learning --- in which partial labels are available. In addition, some theoretical results on affine sparse subspace clustering will be discussed.


Chun-Guang Li is an Associate Professor with the School of Information and Communication Engineering, at Beijing University of Posts and Telecommunications (BUPT). He received his Ph.D. degree in signal and information processing from BUPT in Dec. 2007 and B.E. degree in telecommunication engineering from Jilin University in July 2002. He visited in the Vision, Dynamics and Learning lab at the Center for Imaging Science (CIS) in the Johns Hopkins University (JHU) from Dec. 2012 to Nov. 2013, and the Visual Computing group at Microsoft Research Asia from July 2011 to April 2012. His research interests focus on statistical learning especially for modeling with high dimensional data, including sparse/low-rank model, manifold/subspace clustering, matrix completion, semi-supervised learning and their applications in pattern recognition, computer vision, precise medicine, and etc.

SIST-Seminar 17060