Bayesian Inferences and the Laplace-Bernstein-von Mises Theorem

Add to your list(s) Download to your calendar using vCal

• Professor Richard Nickl
• Friday 20 October 2023, 18:00-19:00
• MR2 in the CMS.

If you have a question about this talk, please contact zl474.

Bayesian inference has been widely used in the statistical sciences and applied mathematics – systematically so since Laplace’s ‘Theorie analytique’ from 1812. It has seen a ‘dark age’ of subjectivist thinking in the 20th century due to computational infeasibility, but then emerged, with the advent of modern Markov chain Monte Carlo methods in the 1990s, as a popular paradigm for data driven inference under uncertainty. Nowadays Bayesian algorithms are among the most commonly used methods in statistics and machine learning. We will discuss some mathematical theorems that explain when one can trust such algorithms in the high-dimensional context of modern data science.

This talk is part of the The Archimedeans series.

This talk is included in these lists:

• MR2 in the CMS
• The Archimedeans
• ob366-ai4er

Note that ex-directory lists are not shown.