“Recent advances in Monte Carlo methods for multi-dimensional signal processing and machine learning”. David Luengo, Victor Elvira , Luca Martino.

Monte Carlo methods are often required in statistical signal processing and machine learning in order to deal with intractable probability density functions. For instance, particle filters, adaptive importance sampling algorithms (like population Monte Carlo (PMC)) and Markov Chain Monte Carlo (MCMC) methods have been widely used within the machine learning and signal processing communities for a long time. Over the last years, many extensions and variants of these families of methods have been developed to improve their performance (e.g., for the estimation of fixed parameters or dealing with multi-modal target densities): particle MCMC (PMCMC), the SMC2 algorithm, adaptive MCMC approaches (i.e., MCMC with adaptive proposal functions), multiple try Metropolis (MTM) strategies, marginal MCMC methods, etc. Some of these methods have found their way into the signal processing and machine learning literature, but there are still many recent advanced Monte Carlo methods that are not so widely known. This special session intends to promote the exchange of ideas between members of the signal processing, machine learning and statistics communities with the aim of discussing recent advances in Monte Carlo methods and their applications.