[edit]

Proceedings of Machine Learning Research Frequently Asked Questions

The Proceedings of Machine Learning Research are very much a community effort. Thanks are particularly due to Isabelle Guyon for working closely with us in preparing early volumes, Nicola Talbot for preparing the LaTeX class file and Leslie Pack Kaelbling for supporting the launch in the first place.

There's a Mistake in a Paper: Can it be Fixed?

If the mistake is a clear typo (e.g. abstract or authors not matching to the PDF) you can submit a pull request to fix it. You'll need a GitHub account for this. Log into your GitHub account and click the "edit" button on the abstract page and if you are logged into GitHub you will be able to edit the meta information directly.

Pages are generated by jekyll in github, which is a static webpage generation framework. Information for each paper is included in the `_posts/` subdirectory, clicking edit from the abstract page will allow you to edit the relevant file. Each file is a markdown with a yaml header. All the meta information is in the yaml header. The format is based on Martin Fenner's citeproc. Note title information and author information is stored twice. Once in a format appropriate for html (e.g. with unicode characters) and once in a format appropriate for LaTeX (e.g. with LaTeX accents and curly brackets to preserve case). find information on the format

Do not edit the bib file, this has no effect on the generated pages or the generated bibfiles.

If the change is more involved then you need to get permission of the proceedings editors. Please report how you got that permission in the pull request (i.e. what the circumstances were, who the individual who gave you permission was). For example, if there's a mistake in the PDF you will need the volume editor's permission to make changes.

How do I publish my proceedings with you?

Email the series editor with your planned event. You need to provide brief CVs of the organizers and an overview of the subject matter. Potential proceedings are judged on the basis of the appropriateness of the material, the perceived quality of the workshop and the track record of the organizers. We expect papers to be peer reviewed, but we are open to innovative mechanisms on how this is done. Please provide a description of your approach.

We expect proceedings requests to come from the editors of the proceedings directly, not through proxies.

I would like to register a Reissue Proceedings with you?

You will need the support of the original conference organizers. When you have this email the series editors with your planned reissue. The reissue proceedings need to be provided in the format detailed here here

What Volume Number will my Proceedings be?

When you are close to finalizing the proceedings, email the series editor. You'll receive you a volume number by allocating from those currently available.

What if I want a Paper Copy?

You are free to organize a hard copy of your proceedings for your workshop, although PMLR does not provide the support for this. Previous editors have made use of Microtome. See the Challenges in Machine Learning series here which is made up of volumes from PMLR. Nicola Talbot has created a LaTeX package for creating books from the proceedings.

Should the Call for Papers be Before or After the Workshop?

This is up to you. You can make the proceedings a pre-proceedings or a post-proceedings. Some conferences, like AISTATS 2010, do the call before the workshop and put the proceedings on line in time for the event. Other will finalize their proceedings after the event.

Overall our advice would be for larger proceedings and more formal conferences, get them up before the event, for smaller proceedings and less formal workshops it can make sense to produce them after the event.

What is the Style File for the Proceedings?

There are two style files, a two column style file (see for example the AISTATS volumes) or a one column style file similar to the standard JMLR style (although you'll need to set the header to read "Proceedings of Machine Learning Research" instead of "Journal of Machine Learning Research"). As of 2010 Nicola Talbot has kindly put together a LaTeX package for these styles. It is available here. We don't support preparation systems other than LaTeX.

Software for Supporting Proceedings Preparation

As of 2012 Nicola Talbot has prepared a Java Application for assisting in preparing a PMLR proceedings. It is available here. Many thanks to Nicola.

What do I Need to Provide for the Proceedings?

You should appoint a publications chair (who could also be one of the editors, but needn't be: for larger conferences it is often a separate role, but for smaller workshops it tends to be an editor). The publications chair will be responsible for compiling the proceedings.

The proceedings are provided in the format detailed here.

As well as the papers in the format described above, you will need to provide permission to publish forms from your authors. Please distribute the form found here to your authors. Have them sign a copy and send you a scanned PDF. The scanned PDF should be named with the convention given above with the suffix Permission, e.g. turner10aPermission.pdf. These files should be placed in a separate directory from the other material called vXpermissions and sent the series editor.

I'm an editor: How do I make changes to a published proceedings?

To do this you'll need to have a github account. Some familiarity with git will help a lot.

For smaller changes, you just need to be logged into your github account, go to the abstract page of the paper you want to change and click on the ‘edit’ button.

You’ll then go straight to the relevant file and you can edit it directly (note that title and author changes need to be made in two fields, there is one normal field and one for bibtex/latex).
For larger changes, you may need to use the more traditional approach to pull requests. You should

"fork the repository" (see here) and then
"clone it" (see here) to your local drive.
Make the changes locally and commit them and push them then
submit a pull request (see here)

All of this allows us to use github’s mechanisms for version control to track changes to the proceedings.