Masterclass in Computational Methods for Large-Scale Bayesian Inference 2016
Professor Mattias Villani
This masterclass deals with computational aspects of Bayesian econometrics and statistics, with a special focus on large datasets and other computationally demanding problems.
A variety of new generation computational tools will be presented, including pseudo-marginal Markov chain Monte Carlo (MCMC), variational methods, Gaussian process optimisation, distributed MCMC, as well as relevant CPU and GPU parallelisation techniques.
It will also cover a variety of interesting computationally demanding models that are widely used currently in machine learning, and should have their place in big data econometrics. In particular, Gaussian process regression and classification, and topic models for textual data, will be covered.
The masterclass is open to anyone interested in Bayesian econometrics. It is complemented by the Melbourne Bayesian Econometrics Workshop taking place on Monday, 4 July 2016.
Attendance to both the workshop and the masterclass is free, although registration is required.
masterclass and workshop registration