Την Τετάρτη 20/11 θα πραγματοποιηθεί διάλεξη
από τον Stephen Vavasis, καθηγητή στο University of Waterloo (Canada).
Ο διακεκριμένος συνάδελφος, που έχουμε την τιμή να φιλοξενούμε
στο Τμήμα Μηχανικών Η/Υ και Πληροφορικής, θα μιλήσει με θέμα
“Fast projection methods for robust separable nonnegative matrix factorization“
Η ομιλία προγραμματίζεται να δοθεί στις 15:00 (αμφιθέατρο Β4, κτήριο Β).
Επισυνάπτεται περίληψη της ομιλίας και σύντομο βιογραφικό του ομιλητή.
Φιλικά
Ε. Γαλλόπουλος
Καθηγητής
ΤΜΗΥΠ
Fast projection methods for robust separable nonnegative matrix factorization
Nonnegative matrix factorization (NMF) has become a widely used tool for analysis of high-dimensional data. Although the NMF problem is NP-hard, certain special cases, such as the case of separable data with noise, are known to be solvable in polynomial time. We propose a very fast successive projection algorithm for this case. Variants of this algorithm have appeared previously in the literature; our principal contribution is to prove that the algorithm is robust against noise and to formulate bounds on the noise tolerance. A second contribution is an analysis of a preconditioning strategy based on semidefinite programming that significantly improves the robustness against noise. We present computational results on artificial data and on a simulated hyperspectral imaging problem.
This talk represents joint work with Nicolas Gillis of University of Mons (Belgium).
About the speaker
Stephen A. Vavasis is a professor in the Department of Combinatorics and Optimization, University of Waterloo, Canada. He received his PhD in Computer Science in 1989 from Stanford University and held the positions of Assistant, Associate and Full Professor of Computer Science at Cornell University from 1989 to 2006. Prior to that, he received his A.B. from Princeton and a Certificate of Advanced Study from Cambridge, both in mathematics.He has also held summer and visiting appointments at Argonne National Laboratory, Bell Labs, Sandia National Laboratories and other U.S. institutions. He is currently taking a year-long sabbatical as a visitor at University of Patras. His research interests include convex optimization and applications of optimization and matrix methods to information retrieval.