[REVIEW]: PySwarms: a research toolkit for Particle Swarm Optimization in Python

whedon commented 7 years ago

Submitting author: @ljvmiranda921 (Lester James Miranda) Repository: https://github.com/ljvmiranda921/pyswarms Version: v0.1.7 Editor: @kyleniemeyer Reviewer: @stsievert Archive: 10.5281/zenodo.1145432

Status

Status badge code:

HTML: <a href="http://joss.theoj.org/papers/235299884212b9223bce909631e3938b"><img src="http://joss.theoj.org/papers/235299884212b9223bce909631e3938b/status.svg"></a>
Markdown: [![status](http://joss.theoj.org/papers/235299884212b9223bce909631e3938b/status.svg)](http://joss.theoj.org/papers/235299884212b9223bce909631e3938b)

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) 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 questions

@stsievert, please carry out your review in this issue by updating the checklist below (please make sure you're logged in to GitHub). The reviewer guidelines are available here: http://joss.theoj.org/about#reviewer_guidelines. Any questions/concerns please let @kyleniemeyer know.

Conflict of interest

[x] As the reviewer I confirm that I have read the JOSS conflict of interest policy and that there are no conflicts of interest for me to review this work.

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 repository url?
[x] License: Does the repository contain a plain-text LICENSE file with the contents of an OSI approved software license? [Yes, MIT]
[x] Version: Does the release version given match the GitHub release (v0.1.7)? [Yes, a prerelease]
[x] Authorship: Has the submitting author (@ljvmiranda921) made major contributions to the software? Does the full list of paper authors seem appropriate and complete?

Functionality

[x] Installation: Does installation proceed as outlined in the documentation? [Yes, worked for me]
[x] Functionality: Have the functional claims of the software been confirmed? [Yes, docs make it clear that classes have parameters]
[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.) [I see no performance claims]

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. [setup.py, requirements_dev.txt]
[x] Example usage: Do the authors include examples of how to use the software (ideally to solve real-world analysis problems). [Yes, in README and examples/]
[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 function of the software can be verified? [Yes, TravisCI]
[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 [Yes, and linked on docs]

Software paper

[x] Authors: Does the paper.md file include a list of authors with their affiliations?
[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] References: Do all archival references that should have a DOI list one (e.g., papers, datasets, software)?

whedon commented 7 years ago

Hello human, I'm @whedon. I'm here to help you with some common editorial tasks for JOSS. @stsievert it looks like you're currently assigned as the reviewer for this paper :tada:.

:star: Important :star:

If you haven't already, you should seriously consider unsubscribing from GitHub notifications for this (https://github.com/openjournals/joss-reviews) repository. As as reviewer, you're probably currently watching this repository which means for GitHub's default behaviour you will receive notifications (emails) for all JOSS reviews 😿

To fix this do the following two things:

Set yourself as 'Not watching' https://github.com/openjournals/joss-reviews:

watching

You may also like to change your default settings for this watching repositories in your GitHub profile here: https://github.com/settings/notifications

notifications

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

@whedon commands

kyleniemeyer commented 7 years ago

👋 @ljvmiranda921 @stsievert @oesteban

stsievert commented 7 years ago

I've looked this over briefly today, and here's some comments:

I didn't see a statement of need that clearly addressed the audience. Is the audience swarm intelligence researchers? Other uses of neural nets that would some other optimization algorithm?
Particle Swarm Optimization is mentioned many times, but I've never heard of it. Is be used in place of sklearn's GridSearch (i.e., is it useful in hyper parameter tuning for machine learning)? I'd like to see a brief description of what (if any?) guarantees PSOs have.
What's the difference in the packages for global min and local min? If minimizing a convex function these two packages will produce the same result. Are you most concerned about non-convex functions?

ljvmiranda921 commented 7 years ago

Hi @stsievert ,

Thank you so much for your comments.

An aside, PSO is a class of search algorithms that can be used to optimize functions even if they're differentiable or not. This library provides implementations of these algorithms that should solve any optimization problem. Of course, the performance will depend on the PSO parameters (you can control if the particles are independent from one another, follows the herd, etc.), and to the type of PSO used in a certain problem.

As for your comments:

The package is intended for swarm int. researchers & students who would like a high-level interface of implementing PSO in their problems. It doesn't have to be a neural nets problem, in this use-case, it demonstrates how the parameters of a neural-net can be optimized through this package. It also doesn't need to be a ML problem, as long as an objective function is defined, then it can be used as input in pyswarms. This can range from performing inverse-kinematics in robotics, or job-shop scheduling (It just so happened that most of my use-cases are related to ML).
Although you can use swarm intelligence algorithms for finding the hyperparameters in a machine learning classifier [1] [2], similar to GridSearch/RandomSearch in sklearn, the GridSearch and RandomSearch in this package is only for finding the "hyperparameters" for the PSO algorithm (a.k.a. the parameters that controls its behaviour). In a vanilla PSO, you have parameters that controls how much the particles are dependent from one another, or how much of their neighbors they're going to follow.
I believe you're referring to GlobalBestPSO and LocalBestPSO? This doesn't necessarily refer to global minima or local minima. These two define the "topology" of the swarm. In GlobalBestPSO, all particles follow a leader, who has the best cost in the current iteration [3]. In LocalBestPSO, the particles are divided into neighbors/clusters, and only looks at the sub-leader in their respective neighborhoods [4]. I'm using the words GlobalBest and LocalBest because it's the one used in literature [3] [4].

Thank you so much, I hope it clears things up! 😄

[1] Dario Floreano, Peter Durr, and Claudio Mattiussi. “Neuroevolution: from architectures to learning”. In: Evolutionary Intelligence (2008). [2] Kenneth Stanley, David B. D’Ambrosio, and Jason Gauci. “A hypercube-based encoding for evolving large-scale neural networks”. In: Artificial Life (2009). [3] J. Kennedy and R.C. Eberhart, “Particle Swarm Optimization,” Proceedings of the IEEE International Joint Conference on Neural Networks, 1995, pp. 1942-1948. [4] Y. Shi and R.C. Eberhart, “A modified particle swarm optimizer,” Proceedings of the IEEE International Conference on Evolutionary Computation, 1998, pp. 69-73.

stsievert commented 7 years ago

The package is intended for swarm int. researchers & students who would like a high-level interface of implementing PSO in their problems.

Thanks – that clears this up. I don't think the above is expressed clearly in your docs. I would like to see the same sentence, maybe with "high-level interface" expanded.

Thank you so much, I hope it clears things up!

It does! Thank you.

ljvmiranda921 commented 7 years ago

Thank you @stsievert . Unfortunately, I'm on a trip so I might be able to handle this on Thursday, (UTC+09 :00).

stsievert commented 7 years ago

Unfortunately, I'm on a trip so I might be able to handle this on Thursday

No worries – take your time!

ljvmiranda921 commented 7 years ago

Hi @stsievert , I've made the changes you've requested both in Paper.md and the README.rst. Hope it's much better now. Thanks a lot!

stsievert commented 7 years ago

Thanks – this looks better! I've commented in https://github.com/ljvmiranda921/pyswarms/issues/56 for some more small edits.

kyleniemeyer commented 7 years ago

@stsievert thanks for your comments so far! I noticed that a few of the checkboxes at the top are still empty, is your review still in progress?

(Also, please do check off that first one about the conflict of interest)

ljvmiranda921 commented 7 years ago

Update: The latest commit resolves ljvmiranda921/pyswarms#56

Hi @stsievert , the latest commit applies the changes needed. 👍

stsievert commented 7 years ago

@kyleniemeyer my review is still in progress. I still need to download and verify the claims in the documentation. Plus I'd like @oesteban's comment too.

ljvmiranda921 commented 7 years ago

Update: The latest commit resolves ljvmiranda921/pyswarms#61

stsievert commented 7 years ago

I've updated the review, and all boxes are checked.

kyleniemeyer commented 7 years ago

Hi @oesteban, have you had a chance to look at the software package and paper, if you are still able to?

ljvmiranda921 commented 7 years ago

Hi @kyleniemeyer , sorry I wasn't able to check this recently. Thank you for the review @stsievert !

ljvmiranda921 commented 7 years ago

Hi @kyleniemeyer ! :) I am just wondering what the status of this would be. Thanks a lot!