Reviewing Guidelines

Timeline

Milestone	Date
Fill out the questionnaire	TBA
Submission deadline	TBA
Reviews assigned	TBA
Reviews due	TBA
Reviews available to authors	TBA
Author private rebuttal period	TBA
Author–Reviewer discussion period	TBA
AC–Reviewer discussion period	TBA
Meta-reviews due	TBA
Decision released	TBA

Tracks

ML4H accepts submissions in two tracks:

• Proceedings (8 pages): polished work with technical sophistication and clear impact in health. Each reviewer is assigned up to 4 submissions, at most 3 of which are Proceedings submissions.
• Findings (4 pages, non-archival): work prioritized for its likelihood to generate good discussion at the symposium.

Area chairs receive 6–10 submissions collectively.

Registration and assignment

Reviewers and area chairs must complete OpenReview registration by the questionnaire deadline and consent to collection, retention, and analysis of participation data by the Association for Health Learning and Inference (AHLI). Papers are matched with reviewers based on declared data modalities and subject area expertise.

Upon receiving assignments, reviewers must immediately report any conflicts of interest and verify that submissions fall within their areas of expertise.

Confidentiality

Reviewers, area chairs, and program chairs must keep everything relating to the review process confidential. Do not share, discuss, or disclose any information related to submissions with anyone, including any LLMs or other AI assistants. You may not use ideas, code, or results from submissions in your own work unless and until they become publicly available and the author has granted the appropriate permission or license covering your intended use. Code submitted for review may not be distributed to anyone, including any LLMs or other AI assistants, and may not be used for any purpose other than the review process. If a reviewer or area chair needs additional support via an external reviewer, they should be formally invited through OpenReview. Reviews and discussions are confidential in the same way submissions are, and you remain fully responsible for the content of your review.

Finally, reviewer and area chair assignments are made to ensure that there are no conflicts of interest. Reviewers and area chairs should not discuss any of their assigned papers with other reviewers or area chairs that are not assigned to the same submission. It is also a violation for reviewers and area chairs to ask other reviewers or area chairs for information about their own submissions as authors. The process is double-blind: authors and reviewers do not know each other’s identities, though assigned reviewers and area chairs can see each other’s identities.

Generative AI Policy

ML4H prohibits the use of generative AI tools, including LLMs and other AI assistants, in reviewing. Privileged information — such as the submissions themselves, reviews of the submissions, and discussions about the submissions — may only be used for the purpose of reviewing. Therefore, we cannot allow this information to be submitted to external services such as generative AI tools.

The reviewing process is meant to obtain reviews from peer researchers who have read, understood, and evaluated the correctness and merits of the submissions. The credibility of ML4H is damaged if reviews are produced using generative AI tools such as LLMs or other AI assistants, whether they are accessed via an external service or on a local machine. This applies to reviews, meta-reviews, and discussion posts, and to all reviewers, area chairs, and program chairs.

Format

Submissions must use official ML4H templates with strict page limits. Reviewers are not obligated to read supplementary materials. Major formatting violations (e.g., a 20-page journal-like submission) are grounds for rejection.

Writing your review

Reviews should provide actionable, constructive, and respectful feedback while helping program leadership identify the highest-impact work. Critique the work, not the authors. Be specific: if something is unclear or incorrect, say exactly what and why.

When evaluating a submission, consider:

• Relevance — Is the work within scope for ML4H?
• Novelty — What is new relative to prior work?
• Technical soundness — Are the methods correct and well-described?
• Experimental rigor — Are baselines appropriate and results reproducible?
• Clinical validity — Does the clinical context justify the approach?
• Clarity — Is the paper well-written and easy to follow?

Discussion phase

During the author–reviewer discussion period, reviewers are expected to engage with one another and their area chair to reach a unified consensus. Area chairs will lead this discussion and flag reviews that lack quality or sufficient engagement with author responses.

Area chair responsibilities

Area chairs should:

1. Check assignments for expertise and conflicts promptly after assignment
2. Flag reviews lacking quality or sufficient responses during the author rebuttal period
3. Lead reviewer discussions during the discussion period
4. Submit meta-reviews by the deadline

Meta-reviews should summarize reviewer opinions, provide salient points for the author, identify areas for improvement, and highlight excellent papers and reviews for potential awards.

Conflicts of interest

Declare any conflicts of interest immediately. Do not review work from authors at your institution, former advisees or advisors, or anyone with whom you have a close personal relationship.