Advanced Machine Learning Techniques for General Nonlinear and Non-Gaussian Problems
Category: Bernoulli Society for Mathematical Statistics and Probability (BS)
Abstract
Sequential Monte Carlo (SMC) as a class of online learning algorithms can handle general non-linear and non-Gaussian modeling and inference. Hence, it has been widely used in signal and image processing, Bayesian inference, risk analysis and rare event sampling, engineering and robotics, bioinformatics, phylogenetics, mathematical finance, etc. This section will present explainable and interpretable SMC algorithms that are suitable for nonlinear and non-Gaussian data science challenges, with rigorous performance guarantees and important scientific applications.