What we'll do
After a long break in meetups (sorry about that!) we're really excited to have Dan Simpson as our speaker for this event. Dan has done lots of really cool Bayesian work and, among other things, is one of the core developers of Stan and has worked on the approximate inference method INLA.
Here's what he's going to talk about:
"Yes, but does it work?"
Abstract: Probably the hardest question for anyone developing or applying a new method to deal with complicated data is "did that actually work?". In this talk, I will go through some of the tricks that can be used when building and deploying Bayesian models to give modellers more confidence that things will be ok. In the end, our most powerful tools are pictures and fear.
And here's a short bio about Dan:
Dan Simpson is the Canadian Research Chair in Bayesian Spatial Modelling and an Assistant Professor at the University of Toronto in Canada. He got his PhD from QUT in 2009 and has since worked in Sweden, Norway, Finland, England, and Canada. His interests include Bayesian modelling, model assessment and checking, visualization, and computing.
Rough time plan:[masked]pm: Arrive, chat, beers
7-8pm: Dan's talk & Questions[masked]pm: More chatting followed by possible restaurant / pub
It should be great fun. Hope to see many of you there!