Quasi‐Monte Carlo sampling method
Date: 17 April 2020 Friday
Date: 17 April 2020 Friday
Time: 3-4pm
Speaker: Dr. Houying Zhu
Abstract: The Monte Carlo method
is one of the widely used numerical methods for simulating probability
distributions by computer‐generated pseudorandom numbers. Quasi‐Monte
Carlo (QMC) methods, which can be seen as a deterministic version of Monte
Carlo methods, have been developed to improve the convergence rate to achieve
greater accuracy, which partially depends on generating samples with a small
discrepancy. In this talk, we consider the role of quasi‐Monte
Carlo idea in statistical sampling and propose the explicit construction of low
discrepancy sequences with respect to non‐uniform
distributions. We also would like to illustrate how to use QMC in practice, for
instance, by integrating QMC rules with Markov Chain Monte Carlo framework for
statistical learning problems such as variable selection.