Workshops

Convex relaxation methods are studied and applied within a variety of disciplines in computer science and mathematics. They aim at providing exact or tight approximations of solutions of difficult problems. In the last few years, they have played a major role in designing efficient algorithms for compressed sensing and level set method. In addition to the substantial impact of convex relaxation methods in applied areas, they also are connected to various branches of mathematical sciences including optimization, functional analysis, geometry, graph theory and combinatorics.

The goal of this workshop is to bring together an interdisciplinary community from mathematics, computer vision, engineering and machine learning to discuss the latest progress and highlight various mathematical questions and algorithmic challenges.

The workshop will discuss the following topics:

1. connections between convex relaxation methods and nonsmooth/nonlinear optimization algorithms based on L^1 and total variation
2. relationships between graph theory, combinatorial and continuous optimizations via convex relaxation techniques
3. relaxation methods to spectral and inference data models in machine learning
4. opportunities of convex relaxation techniques for novel applications in signal processing, image processing, machine learning, computer vision, and graph theory

This workshop will include a poster session; a request for posters will be sent to registered participants in advance of the workshop.

Organizing Committee

Xavier Bresson (City University of Hong Kong)
Antonin Chambolle (École Polytechnique)
Tony Chan (Hong Kong University of Science and Technology)
Daniel Cremers (Technische Universtitat München)
Stanley Osher (University of California, Los Angeles (UCLA))
Thomas Pock (Technische Universität Graz)
Gabriele Steidl (Universität Kaiserslautern)