Prof. Radu Craiu, University of Toronto: Bayesian Computation Strategies for Big Data and Intractable Models Abstract

In data de 8 februarie 2023, ora 12:00, in sala C335, campus FSEGA, va avea loc prezentarea sustinuta de catre dl prof. Radu Craiu, University of Toronto, cu titlul:

Bayesian Computation Strategies for Big Data and Intractable Models

eveniment derulat in cadrul INNODES.

Abstract: Bayesian computation is in double jeopardy due to massive data volumes that make the usual MCMC samplers too computationally expensive and/or models with intractable likelihoods. Even under such adverse conditions, Bayesian data analysis can still be conducted using divide and conquer strategies and/or approximate methods such as Approximate Bayesian Computation (ABC) or Bayesian Synthetic Likelihood (BSL). We discuss some recent work that addresses these challenges, including perturbed MCMC samplers that are used within the ABC and BSL paradigms to significantly accelerate computation while maintaining control on computational efficiency.

Dr. Radu V. Craiu is Professor of Statistical Sciences at the University of Toronto. He studied Mathematics at the University of Bucharest (BS 1995, MS 1996) and received a PhD from the Department of Statistics at The University of Chicago in 2001. His main research interests are in computational methods in statistics, especially, Markov chain Monte Carlo algorithms (MCMC), Bayesian inference, copula models, model selection procedures and statistical genetics. He is currently Contributing Editor for the IMS Bulletin and Associate Editor for the Harvard Data Science Review, Journal of Computational and Graphical Statistics, Statistics Surveys, The Canadian Journal of Statistics, and Computational Statistics and Data Analysis. He received the 2016 CRM-SSC prize, is a Fellow of the Institute of Mathematical Statistics, a Fellow of the American Statistical Association, a Faculty Affiliate of the Vector Institute, and an Elected Member of the International Statistical Institute.

La eveniment se poate participa si online la adresa: