Using PGMax to solve the PMP 2D blind deconvolution exp

[antoine-dedieu]

We use PGMax to solve the binary deconvolution experiment presented in the Section 5.6 of the perturb-and-max-product (https://proceedings.neurips.cc/paper/2021/hash/07b1c04a30f798b5506c1ec5acfb9031-Abstract.html) Neurips 2021 paper.

The example pmp_convor.py computes several samples from the posterior of the model via PMP

This addresses #112

[codecov-commenter]

Codecov Report

Merging #127 (b416f00) into master (ad52b65) will not change coverage.
The diff coverage is 100.00%.

@@            Coverage Diff            @@
##            master      #127   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           10        10           
  Lines          854       854           
=========================================
  Hits           854       854           

Impacted Files	Coverage Δ
pgmax/factors/logical.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update ad52b65...b416f00. Read the comment docs.

[StannisZhou]

Made some comments (mostly on making the notebook more easily understandable).

After you are done, could you also create a colab notebook out of it, and add it to the README page, similar to what we already have?

[StannisZhou]

Also I believe this resolves #112 not issue #11

[StannisZhou]

LGTM! Minor comments. Also don't forget to create the colab and add the link to README