Interdisciplinary Workshop on Inference
Inria et les Mines ParisTech organisent le 12 juin 2012 à l'antenne parisienne du centre Inria Paris-Rocquencourt un workshop intitulé "Interdisciplinary workshop on Inference - Information processing in complex systems with applications to traffic forecasting".
- Date : 12/06/2012
This workshop brings together researchers from statistics, statistical mechanics and machine learning interested in the development of methods and algorithms to process data emerging from complex systems. It is an informal continuation of a preceding meeting (PEPS) held in 2009 in Orsay on similar topics, at the interface between statistical physics and computer science. Some recent convergence of interest between statistical mechanics and machine learning resides in particular in the fact that mean-field methods of statistical mechanics have found their counterpart in terms of powerful message passing algorithms in machine learning. Theoretical questions concerning inverse problem, high dimensional data reduction, compression and restoration will be discussed, to confront complementary viewpoints from the statistics, machine learning and statistical mechanics communities. As a guideline, a special attention will be paid on applications concerning road traffic forecast on large scale networks.
- Thibault Espinasse (ESP, Paul Sabatier University, Toulouse): Whittle's approximation for Gaussian fields indexed by graphs: an extension of the theory of time series for applications to road traffic prediction [Abstract]
- Florent Krzakala (ESPCI, Paris): Optimal compressed sensing and statistical physics [Abstract]
- Jean-Michel Loubes (ESP, Paul Sabatier University, Toulouse): Trafic modelling with processes on graphs
- Victorin Martin (Imara, INRIA): a latent Ising model for real-valued variables inference [Abstract]
- Marc Mézard (Lptms, CNRS, Orsay)
- Yufei Han (CAOR, Mines ParisTech and Imara, INRIA): Towards understanding of global traffic states in large-scale transportation networks [Abstract]
- Kazuyuki Tanaka (Tohoku University): Bayesian image modeling by means of generalized sparse prior and loopy belief propagation [Abstract]
- Muneki Yasuda (Tohoku University): Advanced susceptibility propagation [Abstract]
- Lenka Zdeborová (IPhT, CEA Saclay)