All About that Bayes - The seminar is back

[All About That Bayes]

Dear all,

we are pleased to announce that All About that Bayes seminar is back. It will happen (except for rare events) the second Tuesday of each month at 2 p.m (UTC +1) at Campus Pierre et Marie Curie (Sorbonne Université) or online. If you have a guest and/or wish to organise a mirror session (for people outside of Paris), you can contact Julien Stoehr (stoehr[at]ceremade[dot]dauphine[dot]fr) to check available slots.

The program till December is available online: https://sites.google.com/view/all-about-that-bayes/ The next talk (October 11) will be given by Andrew Gelman (Columbia University).

We are looking forward seeing you all.

The organising team

Prior distribution for causal inference

In Bayesian inference, we must specify a model for the data (a likelihood) and a model for parameters (a prior). Consider two questions:

1. Why is it more complicated to specify the likelihood than the prior?

2. In order to specify the prior, how could can we switch between the theoretical literature (invariance, normality assumption, ...) and the applied literature (experts elicitation, robustness, ...)?

I will discuss those question in the domain of causal inference: prior distributions for causal effects, coefficients of regression and the other parameters in causal models.