[REVIEW]: pytorch-widedeep: A flexible package for multimodal-deep-learning

[editorialbot]

Submitting author: !--author-handle-->@5uperpalo<!--end-author-handle-- (Pavol Mulinka) Repository: https://github.com/jrzaurin/pytorch-widedeep Branch with paper.md (empty if default branch): joss_paper Version: 1.2.0 Editor: !--editor-->@osorensen<!--end-editor-- Reviewers: @siboehm, @makoeppel Archive: 10.5281/zenodo.7908172

Status

Status badge code:

HTML: <a href="https://joss.theoj.org/papers/99999f995d931542cb21ce4f4e8a3dc2"><img src="https://joss.theoj.org/papers/99999f995d931542cb21ce4f4e8a3dc2/status.svg"></a>
Markdown: [![status](https://joss.theoj.org/papers/99999f995d931542cb21ce4f4e8a3dc2/status.svg)](https://joss.theoj.org/papers/99999f995d931542cb21ce4f4e8a3dc2)

Reviewers and authors:

Please avoid lengthy details of difficulties in the review thread. Instead, please create a new issue in the target repository and link to those issues (especially acceptance-blockers) by leaving comments in the review thread below. (For completists: if the target issue tracker is also on GitHub, linking the review thread in the issue or vice versa will create corresponding breadcrumb trails in the link target.)

Reviewer instructions & questions

@siboehm & @dataplayer12 & @makoeppel, your review will be checklist based. Each of you will have a separate checklist that you should update when carrying out your review. First of all you need to run this command in a separate comment to create the checklist:

@editorialbot generate my checklist

The reviewer guidelines are available here: https://joss.readthedocs.io/en/latest/reviewer_guidelines.html. Any questions/concerns please let @osorensen know.

✨ Please start on your review when you are able, and be sure to complete your review in the next six weeks, at the very latest ✨

Checklists

📝 Checklist for @makoeppel

📝 Checklist for @siboehm

[editorialbot]

Hello humans, 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]

:point_right::page_facing_up: Download article proof :page_facing_up: View article proof on GitHub :page_facing_up: :point_left:

[makoeppel]

Review checklist for @makoeppel

Conflict of interest

• [x] I confirm that I have read the JOSS conflict of interest (COI) policy and that: I have no COIs with reviewing this work or that any perceived COIs have been waived by JOSS for the purpose of this review.

Code of Conduct

• [x] I confirm that I read and will adhere to the JOSS code of conduct.

General checks

• [x] Repository: Is the source code for this software available at the https://github.com/jrzaurin/pytorch-widedeep?
• [x] License: Does the repository contain a plain-text LICENSE file with the contents of an OSI approved software license?
• [x] Contribution and authorship: Has the submitting author (@5uperpalo) made major contributions to the software? Does the full list of paper authors seem appropriate and complete?
• [x] Substantial scholarly effort: Does this submission meet the scope eligibility described in the JOSS guidelines
• [x] Data sharing: If the paper contains original data, data are accessible to the reviewers. If the paper contains no original data, please check this item.
• [x] Reproducibility: If the paper contains original results, results are entirely reproducible by reviewers. If the paper contains no original results, please check this item.
• [x] Human and animal research: If the paper contains original data research on humans subjects or animals, does it comply with JOSS's human participants research policy and/or animal research policy? If the paper contains no such data, please check this item.

Functionality

• [x] Installation: Does installation proceed as outlined in the documentation?
• [x] Functionality: Have the functional claims of the software been confirmed?
• [x] Performance: If there are any performance claims of the software, have they been confirmed? (If there are no claims, please check off this item.)

Documentation

• [x] A statement of need: Do the authors clearly state what problems the software is designed to solve and who the target audience is?
• [x] Installation instructions: Is there a clearly-stated list of dependencies? Ideally these should be handled with an automated package management solution.
• [x] Example usage: Do the authors include examples of how to use the software (ideally to solve real-world analysis problems).
• [x] Functionality documentation: Is the core functionality of the software documented to a satisfactory level (e.g., API method documentation)?
• [x] Automated tests: Are there automated tests or manual steps described so that the functionality of the software can be verified?
• [x] Community guidelines: Are there clear guidelines for third parties wishing to 1) Contribute to the software 2) Report issues or problems with the software 3) Seek support

Software paper

• [x] Summary: Has a clear description of the high-level functionality and purpose of the software for a diverse, non-specialist audience been provided?
• [x] A statement of need: Does the paper have a section titled 'Statement of need' that clearly states what problems the software is designed to solve, who the target audience is, and its relation to other work?
• [x] State of the field: Do the authors describe how this software compares to other commonly-used packages?
• [x] Quality of writing: Is the paper well written (i.e., it does not require editing for structure, language, or writing quality)?
• [x] References: Is the list of references complete, and is everything cited appropriately that should be cited (e.g., papers, datasets, software)? Do references in the text use the proper citation syntax?

[makoeppel]

Following line 50 of your paper, can you briefly explain what \sigma and b are? https://github.com/jrzaurin/pytorch-widedeep/issues/118

[5uperpalo]

@makoeppel first of all, thank you for doing the review... from the cited paper(see attached screenshot, @jrzaurin please correct me if I am wrong):

σ(·) is the sigmoid function
b is the bias term

question - should we add the explanation to the JOSS paper? or wait for possible other comments and add it afterwards?

[makoeppel]

Contributing link in the readme gets an page not found error: https://github.com/jrzaurin/pytorch-widedeep/issues/117

[makoeppel]

@editorialbot check references

[editorialbot]

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

OK DOIs

- 10.48550/ARXIV.2003.06505 is OK
- 10.48550/ARXIV.2108.09084 is OK
- 10.18653/v1/N16-1174 is OK
- 10.1162/neco.1997.9.8.1735 is OK

MISSING DOIs

- 10.1109/cvpr.2016.90 may be a valid DOI for title: Deep residual learning for image recognition
- 10.1109/cvpr.2018.00716 may be a valid DOI for title: Shufflenet: An extremely efficient convolutional neural network for mobile devices
- 10.1109/cvpr.2017.634 may be a valid DOI for title: Aggregated residual transformations for deep neural networks
- 10.5244/c.30.87 may be a valid DOI for title: Wide residual networks
- 10.1109/tnnls.2022.3158966 may be a valid DOI for title: RegNet: self-regulated network for image classification
- 10.1109/cvpr.2019.00293 may be a valid DOI for title: Mnasnet: Platform-aware neural architecture search for mobile
- 10.3115/v1/w14-4012 may be a valid DOI for title: On the properties of neural machine translation: Encoder-decoder approaches

INVALID DOIs

- None

[makoeppel]

Can you maybe compare your framework to the Wide & Deep / Deep & Cross models from Keras / Tensorflow? Here are some (incomplete) links to them:

• https://www.tensorflow.org/api_docs/python/tf/keras/experimental/WideDeepModel
• https://keras.io/examples/structured_data/wide_deep_cross_networks/
• https://ai.googleblog.com/2016/06/wide-deep-learning-better-together-with.html?m=1

[osorensen]

@makoeppel first of all, thank you for doing the review... from the cited paper(see attached screenshot, @jrzaurin please correct me if I am wrong):

σ(·) is the sigmoid function
b is the bias term

question - should we add the explanation to the JOSS paper? or wait for possible other comments and add it afterwards?

Nice to see the review has already started!

Regarding your question @5uperpalo, as editor I'd just like the chime in that given the interactive nature of the review process, I think editing this already now sounds like a good idea.

[makoeppel]

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

OK DOIs

- 10.48550/ARXIV.2003.06505 is OK
- 10.48550/ARXIV.2108.09084 is OK
- 10.18653/v1/N16-1174 is OK
- 10.1162/neco.1997.9.8.1735 is OK

MISSING DOIs

- 10.1109/cvpr.2016.90 may be a valid DOI for title: Deep residual learning for image recognition
- 10.1109/cvpr.2018.00716 may be a valid DOI for title: Shufflenet: An extremely efficient convolutional neural network for mobile devices
- 10.1109/cvpr.2017.634 may be a valid DOI for title: Aggregated residual transformations for deep neural networks
- 10.5244/c.30.87 may be a valid DOI for title: Wide residual networks
- 10.1109/tnnls.2022.3158966 may be a valid DOI for title: RegNet: self-regulated network for image classification
- 10.1109/cvpr.2019.00293 may be a valid DOI for title: Mnasnet: Platform-aware neural architecture search for mobile
- 10.3115/v1/w14-4012 may be a valid DOI for title: On the properties of neural machine translation: Encoder-decoder approaches

INVALID DOIs

- None

Hi @5uperpalo, as you can see the bot shows some missing DOIs for the references. Can you check this? Otherwise I am happy with the current status of the paper.md.

[5uperpalo]

@makoeppel , this is strange, I added al of those DOIs after I noticed this message (1-2 weeks ago?), they are all in paper.bib, any idea why is this still showing up?

• https://github.com/jrzaurin/pytorch-widedeep/blob/d05352c73283f6b830225976f6bb7b7699fc658e/mkdocs/paper_JOSS/paper.bib#L177
• https://github.com/jrzaurin/pytorch-widedeep/blob/d05352c73283f6b830225976f6bb7b7699fc658e/mkdocs/paper_JOSS/paper.bib#L187
• https://github.com/jrzaurin/pytorch-widedeep/blob/d05352c73283f6b830225976f6bb7b7699fc658e/mkdocs/paper_JOSS/paper.bib#L196
• https://github.com/jrzaurin/pytorch-widedeep/blob/d05352c73283f6b830225976f6bb7b7699fc658e/mkdocs/paper_JOSS/paper.bib#L205
• https://github.com/jrzaurin/pytorch-widedeep/blob/d05352c73283f6b830225976f6bb7b7699fc658e/mkdocs/paper_JOSS/paper.bib#L213
• https://github.com/jrzaurin/pytorch-widedeep/blob/d05352c73283f6b830225976f6bb7b7699fc658e/mkdocs/paper_JOSS/paper.bib#L237
• https://github.com/jrzaurin/pytorch-widedeep/blob/d05352c73283f6b830225976f6bb7b7699fc658e/mkdocs/paper_JOSS/paper.bib#L273

[makoeppel]

@editorialbot generate pdf

[editorialbot]

:point_right::page_facing_up: Download article proof :page_facing_up: View article proof on GitHub :page_facing_up: :point_left:

[makoeppel]

@editorialbot check references

[editorialbot]

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

OK DOIs

- 10.48550/ARXIV.2003.06505 is OK
- 10.48550/ARXIV.2108.09084 is OK
- 10.48550/arXiv.1606.07792 is OK
- 10.18653/v1/N16-1174 is OK
- 10.1109/cvpr.2016.90 is OK
- 10.1109/CVPR.2018.00716 is OK
- 10.1109/CVPR.2017.634 is OK
- 10.5244/c.30.87 is OK
- 10.1109/TNNLS.2022.3158966 is OK
- 10.1109/CVPR.2019.00293 is OK
- 10.1162/neco.1997.9.8.1735 is OK
- 10.3115/v1/w14-4012 is OK
- 10.48550/arXiv.1708.05123 is OK

MISSING DOIs

- 10.1163/1574-9347_dnp_e612900 may be a valid DOI for title: keras

INVALID DOIs

- None

[makoeppel]

@makoeppel , this is strange, I added al of those DOIs after I noticed this message (1-2 weeks ago?), they are all in paper.bib, any idea why is this still showing up?

* https://github.com/jrzaurin/pytorch-widedeep/blob/d05352c73283f6b830225976f6bb7b7699fc658e/mkdocs/paper_JOSS/paper.bib#L177

* https://github.com/jrzaurin/pytorch-widedeep/blob/d05352c73283f6b830225976f6bb7b7699fc658e/mkdocs/paper_JOSS/paper.bib#L187

* https://github.com/jrzaurin/pytorch-widedeep/blob/d05352c73283f6b830225976f6bb7b7699fc658e/mkdocs/paper_JOSS/paper.bib#L196

* https://github.com/jrzaurin/pytorch-widedeep/blob/d05352c73283f6b830225976f6bb7b7699fc658e/mkdocs/paper_JOSS/paper.bib#L205

* https://github.com/jrzaurin/pytorch-widedeep/blob/d05352c73283f6b830225976f6bb7b7699fc658e/mkdocs/paper_JOSS/paper.bib#L213

* https://github.com/jrzaurin/pytorch-widedeep/blob/d05352c73283f6b830225976f6bb7b7699fc658e/mkdocs/paper_JOSS/paper.bib#L237

* https://github.com/jrzaurin/pytorch-widedeep/blob/d05352c73283f6b830225976f6bb7b7699fc658e/mkdocs/paper_JOSS/paper.bib#L273

@5uperpalo sorry this was on my end I did not notice that you changed them already. However, there is still one missing.

[editorialbot]

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

OK DOIs

- 10.48550/ARXIV.2003.06505 is OK
- 10.48550/ARXIV.2108.09084 is OK
- 10.48550/arXiv.1606.07792 is OK
- 10.18653/v1/N16-1174 is OK
- 10.1109/cvpr.2016.90 is OK
- 10.1109/CVPR.2018.00716 is OK
- 10.1109/CVPR.2017.634 is OK
- 10.5244/c.30.87 is OK
- 10.1109/TNNLS.2022.3158966 is OK
- 10.1109/CVPR.2019.00293 is OK
- 10.1162/neco.1997.9.8.1735 is OK
- 10.3115/v1/w14-4012 is OK
- 10.48550/arXiv.1708.05123 is OK

MISSING DOIs

- 10.1163/1574-9347_dnp_e612900 may be a valid DOI for title: keras

INVALID DOIs

- None

[5uperpalo]

@editorialbot generate pdf

[editorialbot]

:point_right::page_facing_up: Download article proof :page_facing_up: View article proof on GitHub :page_facing_up: :point_left:

[siboehm]

Review checklist for @siboehm

Conflict of interest

• [x] I confirm that I have read the JOSS conflict of interest (COI) policy and that: I have no COIs with reviewing this work or that any perceived COIs have been waived by JOSS for the purpose of this review.

Code of Conduct

• [x] I confirm that I read and will adhere to the JOSS code of conduct.

General checks

• [x] Repository: Is the source code for this software available at the https://github.com/jrzaurin/pytorch-widedeep?
• [x] License: Does the repository contain a plain-text LICENSE file with the contents of an OSI approved software license?
• [x] Contribution and authorship: Has the submitting author (@5uperpalo) made major contributions to the software? Does the full list of paper authors seem appropriate and complete?
• [x] Substantial scholarly effort: Does this submission meet the scope eligibility described in the JOSS guidelines
• [x] Data sharing: If the paper contains original data, data are accessible to the reviewers. If the paper contains no original data, please check this item.
• [x] Reproducibility: If the paper contains original results, results are entirely reproducible by reviewers. If the paper contains no original results, please check this item.
• [x] Human and animal research: If the paper contains original data research on humans subjects or animals, does it comply with JOSS's human participants research policy and/or animal research policy? If the paper contains no such data, please check this item.

Functionality

• [x] Installation: Does installation proceed as outlined in the documentation?
• [x] Functionality: Have the functional claims of the software been confirmed?
• [x] Performance: If there are any performance claims of the software, have they been confirmed? (If there are no claims, please check off this item.)

Documentation

• [x] A statement of need: Do the authors clearly state what problems the software is designed to solve and who the target audience is?
• [x] Installation instructions: Is there a clearly-stated list of dependencies? Ideally these should be handled with an automated package management solution.
• [x] Example usage: Do the authors include examples of how to use the software (ideally to solve real-world analysis problems).
• [x] Functionality documentation: Is the core functionality of the software documented to a satisfactory level (e.g., API method documentation)?
• [x] Automated tests: Are there automated tests or manual steps described so that the functionality of the software can be verified?
• [x] Community guidelines: Are there clear guidelines for third parties wishing to 1) Contribute to the software 2) Report issues or problems with the software 3) Seek support

Software paper

• [x] Summary: Has a clear description of the high-level functionality and purpose of the software for a diverse, non-specialist audience been provided?
• [x] A statement of need: Does the paper have a section titled 'Statement of need' that clearly states what problems the software is designed to solve, who the target audience is, and its relation to other work?
• [x] State of the field: Do the authors describe how this software compares to other commonly-used packages?
• [x] Quality of writing: Is the paper well written (i.e., it does not require editing for structure, language, or writing quality)?
• [x] References: Is the list of references complete, and is everything cited appropriately that should be cited (e.g., papers, datasets, software)? Do references in the text use the proper citation syntax?

[5uperpalo]

Hi @siboehm @makoeppel, I believe we fixed the issues with links in the repository. I know this is non-paid, voluntary work... so I don't want to push as I am thankful for your reviews. But, if you have a couple of minutes, please take a 2nd look if everything is ok :).

[makoeppel]

@editorialbot generate pdf

[editorialbot]

:point_right::page_facing_up: Download article proof :page_facing_up: View article proof on GitHub :page_facing_up: :point_left:

[makoeppel]

@editorialbot check references

[editorialbot]

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

OK DOIs

- 10.48550/ARXIV.2003.06505 is OK
- 10.48550/ARXIV.2108.09084 is OK
- 10.48550/arXiv.1606.07792 is OK
- 10.18653/v1/N16-1174 is OK
- 10.1109/cvpr.2016.90 is OK
- 10.1109/CVPR.2018.00716 is OK
- 10.1109/CVPR.2017.634 is OK
- 10.5244/c.30.87 is OK
- 10.1109/TNNLS.2022.3158966 is OK
- 10.1109/CVPR.2019.00293 is OK
- 10.1162/neco.1997.9.8.1735 is OK
- 10.3115/v1/w14-4012 is OK
- 10.1163/1574-9347_dnp_e612900 is OK
- 10.48550/arXiv.1708.05123 is OK

MISSING DOIs

- None

INVALID DOIs

- None

[makoeppel]

@5uperpalo sorry for the delay, times are busy these weeks. However, I only need to finish the functionality checks of the software. I try to get them done asap. The rest looks great 😊

[siboehm]

I will go through it again on the weekend!

[siboehm]

Went through it, the code itself mostly looks good but I have quite a few issues with the written text in it's current state.

1. Paper is missing a comparison to other similar packages in the field, as already pointed out by @makoeppel.
2. l55: In the formula a_deepimage seems to be missing after W_deepimage. The same formula in the documentation contains an a_deepimage term.
3. There are a lot of problems with wording and spelling in the current version of the paper. These should be fixed, before the paper can be accepted. The quickest way to resolve them is probably to pipe the whole text through Grammarly. Some issues (far from exhaustive!):
   1. Pytorch -> PyTorch
   2. l12: frames -> frameworks
   3. l7: What's the different between a data type and a mode?
   4. The sentence starting in l17 is either misspelled or unclear
   5. The sentence starting in l32 is incomplete
   6. l38: an -> a, plethora -> a plethora
4. In the acknowledgements section, the authors mention a lot of other sources of code. It's not clear to me whether the licences of these other projects have been respected (eg ZILNloss is licenced under Apache, but pytorch-widedeep is licensed under MIT, and there's no lincense info in losses.py). See https://github.com/jrzaurin/pytorch-widedeep/issues/127

[osorensen]

Thanks @siboehm! @5uperpalo, please let us know in this thread when you've fixed the issues pointed out by @siboehm. You can use the command @editorialbot generate pdf to re-compile the pdf once you've updated the text.

[osorensen]

👋 @dataplayer12, could you please update us on how it's going with your review?

[osorensen]

Went through it, the code itself mostly looks good but I have quite a few issues with the written text in it's current state.

1. Paper is missing a comparison to other similar packages in the field, as already pointed out by @makoeppel.
2. l55: In the formula a_deepimage seems to be missing after W_deepimage. The same formula in the documentation contains an a_deepimage term.

3. There are a lot of problems with wording and spelling in the current version of the paper. These should be fixed, before the paper can be accepted. The quickest way to resolve them is probably to pipe the whole text through Grammarly. Some issues (far from exhaustive!):

   1. Pytorch -> PyTorch
   2. l12: frames -> frameworks
   3. l7: What's the different between a data type and a mode?
   4. The sentence starting in l17 is either misspelled or unclear
   5. The sentence starting in l32 is incomplete
   6. l38: an -> a, plethora -> a plethora
4. In the acknowledgements section, the authors mention a lot of other sources of code. It's not clear to me whether the licences of these other projects have been respected (eg ZILNloss is licenced under Apache, but pytorch-widedeep is licensed under MIT, and there's no lincense info in losses.py). See Licences of copied code jrzaurin/pytorch-widedeep#127

@5uperpalo, could you please let us know how it's going addressing these points?

[osorensen]

👋 @dataplayer12, could you please update us on how it's going with your review?

[5uperpalo]

@editorialbot generate pdf

[editorialbot]

:point_right::page_facing_up: Download article proof :page_facing_up: View article proof on GitHub :page_facing_up: :point_left:

[5uperpalo]

@editorialbot generate pdf

[editorialbot]

:point_right::page_facing_up: Download article proof :page_facing_up: View article proof on GitHub :page_facing_up: :point_left:

[5uperpalo]

@editorialbot check references

[editorialbot]

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

OK DOIs

- 10.48550/ARXIV.2003.06505 is OK
- 10.48550/ARXIV.2108.09084 is OK
- 10.48550/arXiv.1606.07792 is OK
- 10.18653/v1/N16-1174 is OK
- 10.1109/cvpr.2016.90 is OK
- 10.1109/CVPR.2018.00716 is OK
- 10.1109/CVPR.2017.634 is OK
- 10.5244/c.30.87 is OK
- 10.1109/TNNLS.2022.3158966 is OK
- 10.1109/CVPR.2019.00293 is OK
- 10.1162/neco.1997.9.8.1735 is OK
- 10.3115/v1/w14-4012 is OK
- 10.1163/1574-9347_dnp_e612900 is OK
- 10.48550/arXiv.1708.05123 is OK

MISSING DOIs

- None

INVALID DOIs

- None

[5uperpalo]

first of all, thank you all for the precious time that you are spending on this... my reaction to the previous comments:

1. "Paper is missing a comparison..." I hope we resolved this in this related issue: https://github.com/jrzaurin/pytorch-widedeep/issues/124
2. "l55: In the formula a_deepimage seems to be missing after W_deepimage." - thank you for noticing, I fixed it
3. "There are a lot of problems with wording and spelling... " thank you for noticing I put it through Grammarly as you suggested + re-read it + fixed minor issues (data mode/type - same meaning, I just used type as is more usual, frame/framework, etc)
4. I put both MIT and APACHE2.0 licenses in the repo and mentioned it in the README.MD and I am doing the same for the main branch; see the links and logic I followed for this licensing issue https://github.com/jrzaurin/pytorch-widedeep/issues/127

@makoeppel @osorensen @siboehm @dataplayer12 I do firmly believe we should/could be approaching the final stage 😃

[dataplayer12]

@osorensen sorry for being behind on this. I will submit my review this weekend.

[osorensen]

Thanks @dataplayer12

[5uperpalo]

@dataplayer12 do you, please, by any chance, have time to look at the repo in the upcoming days?

[osorensen]

Thanks for following up on this @5uperpalo. Looking forward to seeing your review @dataplayer12.

[osorensen]

👋 @dataplayer12, could you please update us on how it's going with your review?

[dataplayer12]

@osorensen I think I have held you guys up long enough. Either I will post a review within 24 hours of this comment, or please go ahead without my review. I am terribly sorry for not handling this professionally.

[osorensen]

Thanks for replying quickly @dataplayer12. I fully understand that reviewers have lots of other tasks to do, so no problem. I'll remove you from the reviewer list, and will keep you in mind if I need a reviewer at a later timepoint.

[osorensen]

@editorialbot generate pdf

[osorensen]

@editorialbot check references

[editorialbot]

:point_right::page_facing_up: Download article proof :page_facing_up: View article proof on GitHub :page_facing_up: :point_left: