Thanks to Shane Walsh for the terrible joke (on the theme of iff = if and only if) in the title.
This week I’m in Banff for the “Validating and Expanding Approximate Bayesian Computation Methods” workshop at the BIRS Conference centre. If you ever have the opportunity to attend a conference here I can thoroughly recommend it: amazing scenery and wildlife (I’ve seen elk wandering the campus; another participant has seen a grey wolf!) and great facilities (all meals and accommodation provided free on site; free pool, gym, climbing wall, etc). The discussion in this particular workshop has been very interesting so far: a key theme being whether or not posterior regression adjustment is a good idea (for sure according to asymptotic theory when the model is well specified; see Li & Fearnhead) or a bad idea (for sure according to asymptotic theory when the model mis-specified; talk by David Frazier).
All the talks will be available online here; mine (reviewing some motivations for ABC in astronomy and epidemiology) is already online here. (On my Mac I have to ‘download’ the video to make it play in my browser; there’s a break in the middle of my talk when we left to take a group photo, feel free to fast forward or else listen to see if anyone is caught making candid statements when they thought the camera was switched off!).