How quickly does the Gibbs sampler converge for logconcave distributions?

octubre 14 @ 1:00 pm - 2:30 pm
Cargando Eventos
Seminario de Probabilidad y Estadística

Resumen

The Gibbs sampler is a widely used Markov Chain Monte Carlo algorithm that iteratively samples from conditional distributions of the target probability measure. Assuming log-concavity, we provide a sharp bound on the convergence rate in terms of relative entropy for the random scan version.

Imparte
PhD. Hugo Lavenant

Assistant Professor-Bocconi University”

Detalles

Fecha:
octubre 14
Hora:
1:00 pm - 2:30 pm
Categoría del Evento:

Organizador

Departamento de Probabilidad y Estadistica

Lugar

Salón 13, Edificio C del IIMAS