[PRE REVIEW]: Zoobot: Adaptable Deep Learning Models for Galaxy Morphology

[editorialbot]

Submitting author: @mwalmsley (Mike Walmsley)
Repository: https://github.com/mwalmsley/zoobot
Branch with paper.md (empty if default branch): docs
Version: v1.0.0
Editor: @plaplant
Reviewers: @crhea93, @devanshkv
Managing EiC: Dan Foreman-Mackey

Status

Status badge code:

HTML: <a href="https://joss.theoj.org/papers/447561ee2de4709eddb704e18bee846f"><img src="https://joss.theoj.org/papers/447561ee2de4709eddb704e18bee846f/status.svg"></a>
Markdown: [![status](https://joss.theoj.org/papers/447561ee2de4709eddb704e18bee846f/status.svg)](https://joss.theoj.org/papers/447561ee2de4709eddb704e18bee846f)

Author instructions

Thanks for submitting your paper to JOSS @mwalmsley. Currently, there isn't a JOSS editor assigned to your paper.

@mwalmsley if you have any suggestions for potential reviewers then please mention them here in this thread (without tagging them with an @). In addition, this list of people have already agreed to review for JOSS and may be suitable for this submission (please start at the bottom of the list).

Editor instructions

The JOSS submission bot @editorialbot is here to help you find and assign reviewers and start the main review. To find out what @editorialbot can do for you type:

@editorialbot commands

[editorialbot]

Hello human, I'm @editorialbot, a robot that can help you with some common editorial tasks.

For a list of things I can do to help you, just type:

@editorialbot commands

For example, to regenerate the paper pdf after making changes in the paper's md or bib files, type:

@editorialbot generate pdf

[editorialbot]

Software report:

github.com/AlDanial/cloc v 1.88  T=0.85 s (147.4 files/s, 16680.2 lines/s)
-------------------------------------------------------------------------------
Language                     files          blank        comment           code
-------------------------------------------------------------------------------
Python                          65           2038           3261           4598
Markdown                         7            130              0            407
reStructuredText                29            344            342            325
Bourne Shell                    10             83            143            208
TeX                              1             23              0            207
Jupyter Notebook                 2              0           1638            158
YAML                             6             15             15            107
DOS Batch                        1              8              1             26
Dockerfile                       1              7              7             18
make                             1              4              7              9
HCL                              1              4              4              6
TOML                             1              0              0              6
-------------------------------------------------------------------------------
SUM:                           125           2656           5418           6075
-------------------------------------------------------------------------------


gitinspector failed to run statistical information for the repository

[editorialbot]

Wordcount for paper.md is 1255

[editorialbot]

⚠️ An error happened when generating the pdf.

[editorialbot]

Reference check summary (note 'MISSING' DOIs are suggestions that need verification):

OK DOIs

- 10.48550/arxiv.2206.11927 is OK
- 10.48550/arXiv.1110.3193 is OK
- 10.1093/mnras/stz2816 is OK
- 10.48550/ARXIV.1603.04467 is OK
- 10.1017/S1743921319008615 is OK
- 10.1038/nature14539 is OK
- 10.48550/ARXIV.2302.05442 is OK
- 10.48550/ARXIV.2104.10972 is OK
- 10.5281/zenodo.4414861 is OK
- 10.5281/zenodo.3828935 is OK
- 10.48550/ARXIV.1711.10604 is OK
- 10.48550/ARXIV.2303.00366 is OK

MISSING DOIs

- 10.1093/mnras/stab2093 may be a valid DOI for title: Galaxy Zoo DECaLS: Detailed Visual Morphology Measurements from Volunteers and Deep Learning for 314,000 Galaxies
- 10.1017/pasa.2022.55 may be a valid DOI for title: The DAWES review 10: The impact of deep learning for the analysis of galaxy surveys

INVALID DOIs

- None

[dfm]

@mwalmsley — Thanks for your submission! It looks like there's a syntax error in the author list of your manuscript. The affiliation for the author "Crisel Suárez" should be a single quoted string, not an array!

[mwalmsley]

@editorialbot generate pdf

[editorialbot]

👉📄 Download article proof 📄 View article proof on GitHub 📄 👈

[mwalmsley]

Hi @dfm, apologies for that - typo has been fixed and paper generates correctly.

[mwalmsley]

@editorialbot check references

[editorialbot]

Reference check summary (note 'MISSING' DOIs are suggestions that need verification):

OK DOIs

- 10.48550/arxiv.2206.11927 is OK
- 10.1093/mnras/stab2093 is OK
- 10.48550/arXiv.1110.3193 is OK
- 10.1093/mnras/stz2816 is OK
- 10.48550/ARXIV.1603.04467 is OK
- 10.1017/S1743921319008615 is OK
- 10.1017/pasa.2022.55 is OK
- 10.1038/nature14539 is OK
- 10.48550/ARXIV.2302.05442 is OK
- 10.48550/ARXIV.2104.10972 is OK
- 10.5281/zenodo.4414861 is OK
- 10.5281/zenodo.3828935 is OK
- 10.48550/ARXIV.1711.10604 is OK
- 10.48550/ARXIV.2303.00366 is OK

MISSING DOIs

- None

INVALID DOIs

- None

[dfm]

@editorialbot query scope

@mwalmsley — I'm going to flag this submission for editorial review so that the editorial board can weigh in on the scope, since we explicitly don't publish pre-trained ML models in JOSS. This process normally takes about a week, and I'll keep you posted. I understand that zoobot also offers features for updating the models for new tasks, but if you have any further context about how it fits into the JOSS definition of research software, please feel free to comment on this thread. Thanks for your patience with this process!

[editorialbot]

Submission flagged for editorial review.

[mwalmsley]

@dfm thank you for the explanation and for your time considering this submission.

I would phrase Zoobot as a package for finetuning (and creating from scratch) deep learning galaxy morphology models, which naturally includes many pretrained models, rather than e.g. a list of pretrained models.

I haven't published in JOSS before but my understanding is that code which extracts knowledge from large datasets is in-scope. I would argue that Zoobot meets this bar. It has already been used to measure galaxy morphology in several published works (including the largest detailed morphology catalog to-date) and runs server-side at the Zooniverse selecting which galaxies Galaxy Zoo volunteers should be shown.

Zoobot includes all the code needed for training morphology models from scratch. This is how Galaxy Zoo uses it and this makes up a substantial portion of the code (see e.g. this user-facing function). Training these models required some significant research progress to e.g. handle the Galaxy Zoo decision tree labels as well as a substantial software engineering effort. Zoobot is our effort to package this work and extend it for future Galaxy Zoo projects (including for Euclid, for which Zoobot is part of the official software pipeline) and indeed any project using a decision tree-esque structure.

Further, Zoobot includes extensive code for finetuning these models. User feedback over the beta made it clear that most external users are primarily interested in finetuning the models to their own data. We therefore refocused the API and docs on finetuning. See e.g. the finetuning API, the finetuning guides, and the finetuning and representations examples. I hope that making finetuning straightforward for astronomers without ML backgrounds will be helpful for our field.

I appreciate that if Zoobot was simply, say, a Dropbox folder full of pretrained models (which of course we do include), that would clearly not be a meaningful software package and not be appropriate for JOSS. I do not think that would be an accurate characterization of Zoobot and I am more than happy to discuss this further with the Board.

[dfm]

Thanks @mwalmsley! The editorial team agrees that this submission is within scope and I'm now looking for an editor to handle the submission. I'll post here as soon as someone is assigned.

[mwalmsley]

Thanks @dfm, great news to end the week with.

[plaplant]

@editorialbot assign @plaplant as editor

[editorialbot]

Assigned! @plaplant is now the editor

[plaplant]

@mwalmsley thanks for your submission to JOSS! I'll be the editor for this submission. I will begin reaching out to potential reviewers. If you have any suggestions for people who might be good fits to review, please let me know!

[mwalmsley]

Great @plaplant, appreciate you taking the time! Let me know if potential reviewers aren't immediately available and I can suggest some.

[plaplant]

@editorialbot add @crhea93 as reviewer

[editorialbot]

@crhea93 added to the reviewers list!

[plaplant]

@editorialbot add @devanshkv as reviewer

[editorialbot]

@devanshkv added to the reviewers list!

[plaplant]

@editorialbot start review

[editorialbot]

OK, I've started the review over in #5312.