Sergio Hernandez

PhD Student Computer Science

29/5/2007

Brief Introduction

What the ??? I'm doing

I'm working with Marcus Frean in on-line Bayesian learning using sequential Monte Carlo methods. I've been looking for parallel adaptive schemes for optimizing the behaviour of standard random walks in the parameters space. The sequential learning method will be used for modelling dynamic environments in Bayesian data fusion problems.

Talks - Papers

Bayesian parameter estimation in underwater acoustics (Semacus 2008 - XIth International Seminar on Acoustics - Chile)
Population-based Particle Filtering (Festival of Doubt -2007)
Online Bayesian detection and tracking of multiple objects (New Zealand Computer Science Research Student Conference 2008)
Sequential Monte Carlo (Industrial Research Limited - 2006)
Sequential Learning for General State Space Models (Festival of Doubt -2006)
Multitarget tracking with Poisson process observations (PSIVT 2007)
Reversible Jump MCMC tutorial

More Activities

Computer music with Csound (coming soon!)