SLRA 2015

 

Low-rank matrix approximations appear in a wide range of applications: signal and image processing, systems and control theory, symbolic-numeric computations, etc. In many cases, the matrices are structured (for example, Hankel/Toeplitz, block-Hankel, Sylvester, quasi-Hankel), and a low-rank approximation preserving the matrix structure is desirable. This problem is known as structured low-rank approximation, or SLRA.

The two-day workshop will feature invited talks by renowned experts in SLRA and related topics, including development of efficient algorithmic approaches to deal with these difficult nonconvex problems, analysis of their convergence and theoretical guarantees, convex relaxations of SLRA, applications of SLRA and relations to other techniques like tensor decompositions.

Participation in the event is free of charge and open to everyone, subject to mandatory registration.

Dates

June 1-2, 2015

Location

GIPSA-lab, 11 rue des Mathématiques, Grenoble Campus BP46, F-38402 SAINT MARTIN D'HERES CEDEX, France.

Confirmed invited speakers

• Jérôme Bolte (Université Toulouse 1 Capitole, France)
• Marcus Carlsson (Lund University, Sweden)
• Bart De Moor (Katholieke Universiteit Leuven, Belgium)
• Nicolas Gillis (Université de Mons, Belgium)
• Nina Golyandina (St. Petersburg State, Russia)
• Jean-Baptiste Hiriart-Urruty (Université Toulouse III - Paul Sabatier, France)
• Ivan Markovsky (Vrije Universiteit Brussel, Belgium)
• Bernard Mourrain (INRIA, Sophia Antipolis, France)
• Pierre-Jean Spaenlehauer (INRIA, Nancy, France)
• Stewart Trickett (Absolute Imaging Inc., Calgary, Canada)
• André Uschmajew (University of Bonn, Germany)
• Anatoly Zhigljavsky (Cardiff University, UK)

Organizers

Laurent Condat
Konstantin Usevich

Supported by

• ERC CHESS, Christian Jutten
• ERC DECODA, Pierre Comon

Contact

slra2015 ∂ gipsa-lab ⊙ fr

Registration

Registration to the workshop is mandatory due to limited number participants. Please fill in the registration form.