Optimization for Imaging and Big DataMS13

A large number of Imaging and Big Data applications can be modeled as large-scale optimization problems. Numerical methods that both guarantee some theoretical properties and give good performance are needed when dealing with those problems. The aim of this minisyposium is to describe some novel techniques for such data-intensive applications, as well as emerging challenges in the area. The speakers shall particularly talk about recent interesting applications of optimization techniques to solve problems arising from imaging science, signal processing, wireless communications, machine learning, data mining, information theory, and statistics.

An Active-Set Approach for Minimization over the Simplex and the l1-Ball
Francesco Rinaldi (University of Padova)
Beyond the worst case convergence analysis of the forward-backward algorithm
Silvia Villa (Politecnico di Milano)
Continuous modularity function for detecting leading communities in networks
Francesco Tudisco (University of Strathclyde)
Distributed learning in large scale networks: from GPS-denied localization to MAP inference in a graphical model
Claudia Soares (LARSyS, Instituto Superior Técnico)
Exact spectral-like gradient method for distributed optimization
Nataša Krejić (University of Novi Sad)
Douglas-Rachford iterations for TV - TGV - and constrained TGV - Denoising
Lena Vestweber (Technische Universität Braunschweig)
Optimization in inverse problems via iterative regularization
Guillaume Garrigos (Ecole Normale Supérieure, CNRS)
Stochastic Approximation Method with Second Order Search Direction
Natasa Krklec Jerinkic (University of Novi Sad)
Margherita Porcelli (University of Firenze)
Francesco Rinaldi (University of Padova)
inverse problems, nonlinear optimization, stochastic processes