Joss paper

[odunbar]

Purpose

Closes #245
Closes #261

Content

To Find generated PDF

• go here: .github/workflows/draft-pdf.yml
• Click on latest commit
• Paper is the downloadable "Artifact"

---

• I have read and checked the items on the review checklist.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (a621a04) 88.09% compared to head (a6db621) 88.08%.
Report is 11 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #242      +/-   ##
==========================================
- Coverage   88.09%   88.08%   -0.01%     
==========================================
  Files           7        8       +1     
  Lines        1134     1167      +33     
==========================================
+ Hits          999     1028      +29     
- Misses        135      139       +4     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[lm2612]

on the ABC part
L65 "these tools use statistical emulators to accelerate sampling from a posterior. ABC learns the likelihood directly from the data"
I could be wrong here but my understanding of ABC was that it doesn't necessarily require an emulator, although looking at gpABC they have the option to use GP to accelerate it. The phrasing "ABC learns the likelihood directly from the data" also confused me slightly as it is normally described as "likelihood free"

Something along the lines of:
"these tools both approximately sample from the posterior distribution. In ABC, this approximation comes from bypassing the likelihood that is usually required in sampling methods, such as MCMC. Instead, the objective of ABC is that model output agrees with data based on a scalar-valued objective. In CES, the approximation comes from learning the parameter-to-data map, then following this it calculates an explicit likelihood and uses exact sampling via MCMC. Some ABC algorithms also make use of statistical emulators to further accelerate sampling (gpABC)."

[odunbar]

This is a fair comment! I have adjusted the paragraph.

[odunbar]

NB there are Julia-wide registry issues that are causing test/failures currently. actions can be re-run but buildkite cannot so is being bypassed for now