[PRE REVIEW]: hdlib: A Python library for designing Vector-Symbolic Architectures

[editorialbot]

Submitting author: !--author-handle-->@cumbof<!--end-author-handle-- (Fabio Cumbo) Repository: https://github.com/cumbof/hdlib Branch with paper.md (empty if default branch): joss Version: 0.1.14 Editor: !--editor-->@danielskatz<!--end-editor-- Reviewers: @mahfuz05062, @anilbey Managing EiC: Daniel S. Katz

Status

Status badge code:

HTML: <a href="https://joss.theoj.org/papers/ef404bdbfd9428d01d2142820ab799c1"><img src="https://joss.theoj.org/papers/ef404bdbfd9428d01d2142820ab799c1/status.svg"></a>
Markdown: [![status](https://joss.theoj.org/papers/ef404bdbfd9428d01d2142820ab799c1/status.svg)](https://joss.theoj.org/papers/ef404bdbfd9428d01d2142820ab799c1)

Author instructions

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

@cumbof if you have any suggestions for potential reviewers then please mention them here in this thread (without tagging them with an @). You can search the list of people that have already agreed to review and may be suitable for this submission.

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=0.03 s (437.6 files/s, 118377.4 lines/s)
-------------------------------------------------------------------------------
Language                     files          blank        comment           code
-------------------------------------------------------------------------------
Python                           8            702           1228           1162
TeX                              1             19              0            181
Markdown                         3             82              0            140
Bourne Shell                     1              1              0              2
-------------------------------------------------------------------------------
SUM:                            13            804           1228           1485
-------------------------------------------------------------------------------

gitinspector failed to run statistical information for the repository

[editorialbot]

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

OK DOIs

- 10.3390/a13090233 is OK
- 10.1007/978-3-030-59028-4_1 is OK
- 10.1109/DAC.2018.8465708 is OK
- 10.1007/s12559-009-9009-8 is OK
- 10.1007/s10462-021-10110-3 is OK
- 10.1109/TNNLS.2022.3211274 is OK
- 10.1109/IJCNN55064.2022.9892981 is OK
- 10.1145/3508352.3549477 is OK
- 10.48550/arXiv.2303.15604 is OK
- 10.1109/BIBM55620.2022.9995708 is OK
- 10.1109/DAC18074.2021.9586253 is OK
- 10.1109/FCCM53951.2022.9786145 is OK
- 10.23919/DATE48585.2020.9116397 is OK
- 10.1109/BIBE.2018.00046 is OK
- 10.1007/s12559-021-09974-y is OK
- 10.1109/ACCESS.2020.3001765 is OK
- 10.1109/ALLERTON.2014.7028470 is OK
- 10.1007/s13218-019-00623-z is OK
- 10.48550/arXiv.cs/0412059 is OK
- 10.1016/j.future.2020.04.005 is OK

MISSING DOIs

- None

INVALID DOIs

- None

[editorialbot]

Wordcount for paper.md is 1251

[editorialbot]

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

[danielskatz]

👋 @cumbof - thanks for this submission. Can I suggest that you add a bit to the README about what the software is for, similar to some of the paper introduction?

I'm also going to ask JOSS editors to review this for scope, due to the relatively small amount of code and the short commit history. The editors will discuss if it meets the substantial scholarly effort criterion for review by JOSS. You should hear back in a week or so.

[danielskatz]

@editorialbot query scope

[editorialbot]

Submission flagged for editorial review.

[cumbof]

Thanks @danielskatz for processing this submission so quickly! I'm going to update the README pretty soon (in the next few hours).

Just want to highlight the point that the idea of building this library started from an attempt of generalising chopin2, a software that implements a supervised classification model based on the hyperdimensional computing (HD) paradigm. It also implements a backward variable selection technique as the very first attempt of building a feature selection method with a HD approach.

We generalised everything and we also implemented a series of abstractions that help building vector-symbolic architectures very easily. With this library, we have been able to reimplement chopin2 from scratch drastically improving its running time (see https://github.com/cumbof/hdlib/blob/main/examples/chopin2.py). We are planning to release a new version of chopin2 that makes use of hdlib pretty soon.

Hope in a positive response from the JOSS Editors

[cumbof]

@danielskatz I've updated the README.md as you suggested. Waiting for the Editors evaluation.

[cumbof]

@danielskatz wondering if there is any news about this.

Also note that hdlib has been bumped to version 0.1.11 in the meantime. At the time of the submission last week, its version was 0.1.8. Wondering if this info must also be updated somehow.

[danielskatz]

@editorialbot set v0.1.11 as version

No, @cumbuf, there isn't a decision yet - this is taking a bit longer than normal

[editorialbot]

Done! version is now v0.1.11

[cumbof]

I'll keep waiting. Thanks for your quick reply @danielskatz

[cumbof]

@editorialbot set 0.1.12 as version

@danielskatz hdlib version is now 0.1.12 and, based on my cloc, the number of Python lines (code) in the main branch is 1,272.

Now, since the JOSS documentation reports that:

Total lines of code (LOC). Submissions under 1000 LOC will usually be flagged, thos under 300 LOC will be desk rejected.

Do you still think that asking the Editors to review this submission for the scope is still necessary?

[editorialbot]

I'm sorry @cumbof, I'm afraid I can't do that. That's something only editors are allowed to do.

[danielskatz]

@editorialbot set 0.1.12 as version

yes, let's continue the discussion among the editors

[editorialbot]

Done! version is now 0.1.12

[danielskatz]

@cumbof - the editorial discussion is complete, and we will review this submission

[danielskatz]

@editorialbot assign me as editor

I'll go ahead and edit this, as the most appropriate person with availability

[editorialbot]

Assigned! @danielskatz is now the editor

[danielskatz]

Please give me some suggestions for potential reviewers by mentioning them here in this thread (without tagging them with an @). You can search the list of people that have already agreed to review and may be suitable for this submission, but suggestions don't need to be limited to people on this list, though you should be sure they aren't conflicted with you or this work.

[cumbof]

That's a great news, thanks @danielskatz and JOSS Editors for considering our submission! I'll suggest a few potential reviewers asap.

[cumbof]

@danielskatz here is a list of a few potential reviewers:

• Mohsen Imani (moimani);
• Abbas Rahimi (abbas-rahimi);
• Ross Gayler (rgayler);
• Anıl Tuncel (anilbey);
• Mahfuzur Rahman (mahfuz05062);
• Ritika Giri (ritika-giri).

I would strongly recommend the first three as experts in the field of hyperdimensional computing (a.k.a. vector-symbolic architectures) and the other three as per their strong expertise with Python.

Please note that the first three are not in the list of people that have already agreed to review.

[danielskatz]

👋 @anibey, @mahfuz05062, @rgayler - would one of you be willing to review this submission for JOSS?

[danielskatz]

👋 @abbas-rahimi, @rgayler - Would one of you be willing to review this submission for the Journal of Open Source Software (JOSS)? We carry out our checklist-driven reviews here in GitHub issues and follow these guidelines: https://joss.readthedocs.io/en/latest/review_criteria.html

[danielskatz]

👋 @moimani - Would you be willing to review this submission for the Journal of Open Source Software (JOSS)? We carry out our checklist-driven reviews here in GitHub issues and follow these guidelines: https://joss.readthedocs.io/en/latest/review_criteria.html

[mahfuz05062]

@danielskatz I would be happy to review this submission. Please let me know the process to do that. Thanks!

[rgayler]

@danielskatz thanks for the invitation to review, but I must refuse. Possibly uniquely among those who know about hyperdimensional computing I have never used Python, so am probably uniquely unqualified to review hdlib as software.

While reading the reviewing guidelines in order to decide whether I could adequately review the paper I noticed that the guidelines require related software to be cited. The authors (@cumbof ) should look at the following software to determine the degree of overlap:

• https://github.com/hyperdimensional-computing/torchhd
• https://github.com/Adam-Vandervorst/PyBHV
• https://github.com/UCSD-SEELab/openhd
• https://pypi.org/project/hdtorch/

[danielskatz]

thanks @rgayler

[cumbof]

Thanks for pointing that out @rgayler. I'm not sure I'm allowed to change anything in the manuscript at this point. I'll be waiting for comments from @danielskatz and reviewers

[danielskatz]

@mahfuz05062 - I'll add you to the system as a reviewer, but we won't actually start the review until we get at least one more reviewer as well

[danielskatz]

@editorialbot add @mahfuz05062 as reviewer

[editorialbot]

@mahfuz05062 added to the reviewers list!

[danielskatz]

@cumbof

Thanks for pointing that out @rgayler. I'm not sure I'm allowed to change anything in the manuscript at this point. I'll be waiting for comments from @danielskatz and reviewers

You can make changes before we start the review, so if you can do this in a day or two, please do. And in fact, the review is fairly iterative and this can happen at a detailed level, so it's usually fine to make changes during the review process as well, as long as reviewers can see what you have done (you say so as a comment in this or the review thread, once we get there)

[cumbof]

Sounds great, I'm going to make a few minor changes to the text later today. Thanks @danielskatz

[cumbof]

@danielskatz I've updated the manuscript by adding three references and a couple of short sentences. Also note that a few days ago I bumped the version of the library to 0.1.13

[danielskatz]

@editorialbot set 0.1.13 as version

[editorialbot]

Done! version is now 0.1.13

[danielskatz]

@editorialbot generate pdf

(@cumbof - note that you can do this too when you make changes to the paper source)

[editorialbot]

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

[cumbof]

@danielskatz I've just fixed a couple of minor bugs and bumped the version of hdlib to 0.1.14 There are no other known issues, so this is going to be the version that must be reviewed.

[danielskatz]

@editorialbot set 0.1.14 as version

[editorialbot]

Done! version is now 0.1.14

[danielskatz]

👋 @ritika-giri, @Anibey, @mahfuz05062 - would one of you be willing to review this submission for JOSS?

[mahfuz05062]

I think I already showed my interest in reviewing this earlier.

On Thu, Jul 27, 2023 at 9:11 AM Daniel S. Katz @.***> wrote:

👋 @ritika-giri https://github.com/ritika-giri, @Anibey https://github.com/Anibey, @mahfuz05062 https://github.com/mahfuz05062

• would one of you be willing to review this submission for JOSS?

— Reply to this email directly, view it on GitHub https://github.com/openjournals/joss-reviews/issues/5617#issuecomment-1653706146, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACZD5H6IIOJMVJ4DXYNJBP3XSJZKXANCNFSM6AAAAAAZ4XMMJA . You are receiving this because you were mentioned.Message ID: @.***>

-- Mahfuzur Rahman Ph.D., Computer Science, Sr Software/ML Engineer, Lowe's Lead ML Engineer (Voluntary), Omdena https://www.linkedin.com/in/mahfuzur-rahman-ahm/

[danielskatz]

@mahfuz05062 - yes, sorry, I got a little lost in the issue....

[cumbof]

👋 @ritika-giri, @Anibey, @mahfuz05062 - would one of you be willing to review this submission for JOSS?

@danielskatz wrong mention? You were supposed to mention @anilbey I guess (missing "l")

[ritika-giri]

Hello @danielskatz, unfortunately I will have to decline the review invite at this time.

[danielskatz]

👋 @Anilbey - would you be willing to review this submission for JOSS?