CinC preprint instructions and templates for the 2026 Challenge

[PhysioNet Challenge]

Dear Challengers,

Please see this page for information about preparing your CinC proceedings papers, including instructions, templates, and a checklist:

https://physionetchallenges.org/2026/papers/

Direct links to these documents:

• LaTeX template: https://www.overleaf.com/read/jnnbdgrvcgdb#17231d
• Word template: https://moody-challenge.physionet.org/2026/papers/2026_cinc_template.docx
• Checklist: https://moody-challenge.physionet.org/2026/papers/cinc_proceedings_paper_checklist_for_the_physionet_challenge_2026.pdf

These templates and checklist include important guidance for how to describe your approach and contributions to this year’s Challenge, so please follow the instructions carefully. In particular, please accurately describe your results as indicated in the template, and please cite the data and official description of the Challenge as given on the above page and in the template. These papers describe this year’s Challenge so that you can focus on describing your work; we would much rather you focus on your individual contributions than to repeat a description of the Challenge.

Every year, we need to contact a large number of teams to ask them to update their papers to follow these instructions, and teams that cannot ultimately follow the instructions will not appear in the final conference proceedings. In particular, you should include your scores on the training and validation data in your preprint and presentation. Since you will not receive scores on the test data before you submit your preprint, you must update your preprint after CinC to include your scores and rankings on the test data as well.

These pages from the conference organizers have important information about preparing (and eventually submitting) your preprints and presentations:

https://cinc.org/inf_authors/

https://physionetchallenges.org/2026/#rules

Best,

Audrey

[PhysioNet Challenge]

Dear Challengers,

Due to our late posting of the CinC proceedings papers templates and checklist, we have extended the preprint submission deadline to 1st September 2026, which is the same as the CinC preprint submission deadline. Please use the templates and checklist to carefully prepare and review your preprint, and please be responsive to any feedback about your preprint in the week following the deadline. We require that teams upload their preprints by the deadline for rankings and prize eligibility to support the dissemination and discussion of their work at CinC, and we again look forward to seeing you in Madrid and hearing about your work there in a few short weeks.

Best,
Matt
(On behalf of the Challenge team.)

Please post questions and comments in the forum. However, if your question reveals information about your entry, then please email info at physionetchallenge.org. We may post parts of our reply publicly if we feel that all Challengers should benefit from it. We will not answer emails about the Challenge to any other address. This email is maintained by a group. Please do not email us individually.

[Karen Wurzel]

Greetings Challenge team,

Could you clarify what the requirements are for the mandatory results table? The results tables in the LaTeX and Word templates are very different. Also, the LaTeX table includes evaluations that are not included in either the submission results email or the leaderboard, such as per-site scores.

Best regards,

Karen 

[PhysioNet Challenge]

Dear Karen,

I’m glad you asked, because the tables in both the LaTeX and Word templates needed to be updated. (The version in the Word template was left over from a previous year.) I pushed changes to these. Please use the current version of both templates.

• We do want to include those extra scores, so please run per-site cross-validation on the training set using the same code.
• Also, please include the standard deviation in the per-site cross-validation scores.

Best,

Audrey

[Rodrigo Lozano Puñet]

Dear Challenge team,

Could you please clarify how the per-site cross-validation should be performed?

If no predefined or deterministic folds are provided, different teams will obtain different results depending on how the folds are generated. Reporting the mean and standard deviation does not fully solve this, since the procedure itself still needs to be defined.

In particular, could you clarify:

• How should the folds be generated?
• Should this be leave-one-site-out cross-validation, or a standard k-fold procedure?
• If k-fold is expected, how many folds should be used (e.g., 5, 10, 20)?
• Should we use a fixed random seed or a predefined split?

Given that the submission deadline remains today and these reporting requirements have only just been clarified, it is very difficult to produce reproducible results and comply with the requested format before the deadline.

Please provide a precise definition of the expected evaluation procedure.

Best regards,
Rodrigo

[nutan sahoo]

Thank you for posting this. 

One clarification question is, what do you mean by the validation data? Is that the data used for the scores on the leaderboard? Because Kaggle did not include a validation data. 

On Friday, August 28, 2026 at 3:28:46 PM UTC-4 PhysioNet Challenge wrote:

[PhysioNet Challenge]

Dear nutan884,

Yes, the validation set is the data used for the scores on the leaderboard.

Best,

Audrey

[Arshia Ilaty]

Dear Organizers,

I was wondering whether hackathon was organized as part of this year's challenge. If such an event exists, I would appreciate any information.

Thank you for your time and for organizing this year's challenge.

Best regards,

Ash

[PhysioNet Challenge]

Dear Ash,

There's no hackathon this year due to logistics, but the Sunday Symposium and the rest of the conference should provide ample opportunity for discussion. We look forward to seeing many of you in a little more than a week.

Best,

Matt

(On behalf of the Challenge team.)

Please post questions and comments in the forum. However, if your question reveals information about your entry, then please email info at physionetchallenge.org. We may post parts of our reply publicly if we feel that all Challengers should benefit from it. We will not answer emails about the Challenge to any other address. This email is maintained by a group. Please do not email us individually.