TMLR is now accepting submissions

Raia Hadsell, Kyunghyun Cho, Hugo Larochelle

24 March 2022

As Editors-in-Chief, Hugo, Kyunghyun, and myself are delighted to announce that the Transactions on Machine Learning Research (TMLR) is now accepting submissions! Some of you have expressed your eagerness to send us your manuscripts during the last 3 months since TMLR was announced last NeurIPS, and the reward for your patience is that we are now fully prepared: the TMLR website has submission guidelines and updated LaTeX templates, the OpenReview site is tested and ready, and an expert slate of Action Editors and reviewers have been recruited. It’s exciting to see TMLR becoming a reality, and we are grateful for all our supporters: the volunteers that have helped give feedback and debug the review process; all of you that have agreed to be AEs or reviewers; our advisors and sponsors; and everyone that has expressed enthusiasm for the project.

As it grows, we hope that TMLR will become a central resource for the machine learning community–a continuous thoroughfare of publications and new ideas. TMLR will not have annual issues or proceedings, so we suggest that you keep up to date on new submissions, new accepted papers, and new certifications by signing up for our mailing list digest (either daily or weekly) or following our twitter feeds for general announcements, new submissions, accepted papers and certified papers. If you were not invited to be a reviewer, or were invited but couldn’t accept because of other commitments, please keep an eye out for new submissions that are in your area of expertise - you are welcome to contribute and solicit to review a submission, subject to approval by the action editor.

Looking forward, we hope for a smooth and successful launch of the review process, but of course we are prepared for bumps in the road. In particular, while we have recruited a sizeable pool of over 150 Action Editors and over 1500 reviewers, this will allow us to handle only 150 submissions per month because we have strict quotas on how many papers may be assigned to any AE or reviewer. If this number of submissions is exceeded, it would result in longer review turnarounds or higher quotas, which we have decided is not acceptable. Therefore, in the event that our capacity is reached, TMLR will stop accepting submissions temporarily. We will also be closely monitoring authors with high numbers of submissions, in case it becomes necessary to impose individual limits to maintain fair access for all.

As the first submissions come through, you will start to see different certifications awarded to accepted papers - Survey certifications for papers that include exceptionally thorough or insightful surveys of literature and methods; Reproducibility certifications for papers that go beyond simple verification of a previous result and contribute valuable analyses or additional experiments which provide new insights; Featured certifications for high quality papers which are selected by the Action Editor - roughly the equivalent of choosing a paper for an oral/spotlight at a top tier conference. There will be more certifications over time: the Outstanding certification is awarded by the Editors-In-Chief, approximately on an annual basis, and the Expert Reviewer certification will be awarded to papers authored by very highly rated reviewers, but this will have to wait until we have enough evidence to identify those expert reviewers.

In the future, we hope to use TMLR certifications for reviewing or publishing special topics or collections, for instance through aligning with a workshop or small conference. The certification would be edited and awarded by a guest editor or committee. If you are interested in working with TMLR on developing a certification, please reach out to us.

Another development we are considering for the future is to support non-PDF, interactive submissions. Since TMLR has no paper proceedings, is entirely online and isn’t bound to paper, we see this as an opportunity to innovate on the format of scientific publications. Our first step is to set up an internal committee tasked with exploring our options and making recommendations adapted to TMLR. Stay tuned for more on this!

Of course, there may be other rocky areas to navigate in the first few months! We apologize in advance for any issues that come up; do remember that this is a new effort, and be kind to us. Please check the FAQ regularly whether you are an author, a reviewer, or an Action Editor, and contact us if you encounter any issues or have feedback on the journal.

Raia Hadsell, with
Kyunghyun Cho
Hugo Larochelle