If Bayesian inference is rational, what is approximate inference?

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

• Sebastian Farquhar
• Tuesday 26 April 2022, 15:00-16:00
• SS03, Computer Lab.

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

https://us04web.zoom.us/j/78726354602?pwd=wd4oRhKPoa0PAI00vbr_SlqqV97Mby.1

There are some good reasons to like Bayesian inference, but these are awkwardly mostly also reasons to dislike approximate Bayesian inference. Nevertheless, we need approximations. So are there any ways to understand what would make one approximate inference scheme better than another? I argue that this is very hard. Sequential learning problems like active and continual learning could be seen as evaluations of approximate inference, but doing this carefully is more subtle than it at first seems.

This talk is part of the ML@CL Seminar Series series.

This talk is included in these lists:

• Hanchen DaDaDash
• ML@CL Seminar Series
• SS03, Computer Lab

Note that ex-directory lists are not shown.