MUD-Examples

Examples for Existence, Uniqueness, and Convergence of Parameter Estimates with Maximal Updated Densities

Authors: Troy Butler & Michael Pilosov

Installation

For Python 3.7-3.12:

pip install mud-examples

To reproduce the results in Michael's thesis, use mud-examples==0.1. However, this comes with mud==0.0.28. Newer versions should still produce the same figures.

TeX is recommended (but not required):

apt-get install -yqq \
    texlive-base \
    texlive-latex-base \
    texlive-latex-extra \
    texlive-fonts-recommended \
    texlive-fonts-extra \
    texlive-science \
    latexmk \
    dvipng \
    cm-super

Quickstart

Generate all of the figures the way they are referenced in the paper:

mud_run_all

The above is equivalent to running all of the examples sequentially:

mud_run_inv
mud_run_lin
mud_run_ode
mud_run_pde

Usage

The mud_run_X scripts all call the same primary entrypoint, which you can call with the console script mud_examples.

Here are two examples:

mud_examples --example ode

mud_examples --example lin

and so on. (More on this later, once argparsing is better handled, they might just be entrypoints to the modules themselves rather than a central runner.py, which really only exists to compare several experiments, so perhaps it warrants renaming to reflect that).