MahNMF: Manhattan Non-negative Matrix Factorization
Dacheng Tao ( University of Technology, Sydney)
NICTA SML SEMINARDATE: 2012-09-06
TIME: 11:15:00 - 12:00:00
LOCATION: NICTA - 7 London Circuit
CONTACT: JavaScript must be enabled to display this email address.
ABSTRACT:
Non-negative matrix factorization (NMF) approximates a non-negative matrix X by the product of two non-negative low-rank factor matrices W and H. We introduce Manhattan NMF (MahNMF) to minimize the Manhattan distance between X and W^TH for modeling the heavy tailed Laplace noise. Similar to sparse and low-rank matrix decomposition models, such as robust principal component analysis and GoDec, MahNMF robustly estimates the low-rank part and the sparse part of a non-negative matrix and thus performs effectively when data are contaminated by outliers. MahNMF can be solved by two fast optimization algorithms: rank-one residual iteration and Nesterovas smoothing method, and is applicable to various practical problems by considering different constraints and regularizations, such as box-constraint, manifold regularization, and group sparsity. Experiments are conducted on both synthetic and real-world datasets, such as face images, natural scene images, surveillance videos and multi-view natural image datasets.
BIO:
Dacheng Tao is Professor of Computer Science with the Centre for Quantum Computation & Intelligent Systems and the Faculty of Engineering & Information Technology in the University of Technology, Sydney.
He mainly applies statistics and mathematics for data analysis problems in machine learning, data mining, computer vision, image processing, and video surveillance. He has authored 200+ scientific articles at top venues including IEEE T-PAMI, T-IP, T-NNLS, CVPR, ECCV, AISTATS, ICDM; ACM Multimedia and KDD, with best paper awards, such as the best theory/algorithm paper runner up award in IEEE ICDMa07. He holds the K. C. WONG Education Foundation Award through the National Lab of Pattern Recognition in the Chinese Academy of Sciences. He serves and served as an associate editor for IEEE Transactions on Systems, Man and Cybernetics: Part B (T-SMCB), IEEE Transactions on Circuits and Systems for Video Technology (T-CSVT), IEEE Transactions on Knowledge and Data Engineering (T-KDE), Pattern Recognition (Elsevier), Computational Statistics & Data Analysis (Elsevier), and Signal Processing (Elsevier). He has (co-)chaired for special sessions, invited sessions, workshops, panels and conferences. He is a senior member of the IEEE and a Fellow of the IAPR.
