[HACK] Global vs Local summaries

[bwandelt]

Global vs Local summaries

Compare locally optimal summaries (like those produced by IMNN, using Fisher information) with globally optimal summaries (using mutual information between summaries and parameters over the entire parameter space) for slightly challenging test cases (e.g. very non-Gaussian posterior)

Contacts: Benjamin Wandelt
Participants:

Goals and deliverable

Compare training difficulty and performance of locally and globally optimal summaries

Resources needed

A plot comparing IMNN and IMNN using mutual info for a non-Gaussian test example (heavy-tailed, or multimodal)

Detailed description

[EiffL]

3 types of situations:

• heavy tail
• multimodal
• non trivial (like some kind of circle, polarisation)

Some thoughts:

• Estimating frequencies, even 1D