As a reviewer you are central to the program creation process for CoRL 2026. Your Area Chairs (ACs), Senior Area Chairs (SACs), and the Program Chairs (PCs) will rely greatly on your expertise and your diligent and thorough reviews to make decisions on each paper. Therefore, your role as a reviewer is critical to ensuring a strong program for CoRL 2026.
High-quality reviews are also very valuable for helping authors improve their work, whether it is eventually accepted by CoRL 2026, or not. Therefore it is important to treat each valid CoRL 2026 submission with equal care.
As a token of our appreciation for your essential work, top reviewers will be acknowledged permanently on the CoRL 2026 website.
New to reviewing for CoRL?
Welcome, and thank you for taking this role. A quick orientation on what CoRL papers are judged on. CoRL is a selective, top-tier venue for robot learning, so the first question for any submission is relevance to robot learning: does it address a question that is relevant to learning for physical robots, and does it evaluate the proposed solution convincingly?
Beyond that, weigh significance and novelty, whether the claims are objectively established by the experiments or analysis, and the potential for scientific or technological impact. Wherever a method targets physical robots, evaluation on real hardware is the strongest evidence you can ask for, and a credible sim-to-real demonstration is the next best thing; reward papers that close the loop on a real platform and be appropriately skeptical of strong real-world claims supported only by simulation. At the same time, apply this fairly rather than mechanically. Simulation-only results can be acceptable when the simulation is realistic and the relevance to real robot learning is clear (for example, autonomous-driving results in a realistic simulator), and CoRL also welcomes theoretical contributions and work where simulation is the appropriate testbed. The point is to value real-robot evidence highly and to ask, when it is missing, whether the paper still credibly speaks to learning on physical robots through transfer, data efficiency, or sound analysis. Also check that the required Limitations section honestly discusses assumptions and failure modes; a frank limitations section is a sign of a careful paper, not a weakness.
On the process: read each paper and form your own assessment before doing anything else, since the judgment and the score must be yours even if AI tools or discussion refine your view later. Write the review section by section using the form, keeping weaknesses specific and constructive. Score against the rubric rather than against the other papers in your batch: weak accept (5) is the acceptance threshold, so calibrate deliberately and do not inflate or deflate, and set your confidence score honestly so the Area Chair knows how to weigh your review. Two things carry extra weight this year. First, because of the first-round rejection policy, your initial score can decide whether a paper even reaches rebuttal, so before scoring below weak accept ask whether the issues are rebuttable or reflect deeper problems, and make sure your written review can stand on its own if the authors never get to respond. Second, the post-rebuttal box is required: acknowledge that you read the rebuttal, say whether it changed your view, update your score if so, and justify your final recommendation. The July 13 reviewing deadline is firm; if illness, travel, or a late-discovered conflict puts it at risk, tell your Area Chair as early as possible. Declare any conflict of interest the moment you spot one, and keep submissions strictly confidential: never upload the manuscript or any part of it to AI services that lack enterprise-level confidentiality guarantees. Your Area Chair is your point of contact for anything you are unsure about.
Submissions will be evaluated based on the significance and novelty of the results, either theoretical or empirical. Results will be judged on the degree to which they have been objectively established and/or their potential for scientific and technological impact, as well as their relevance to robotic learning. Authors of regular paper submissions will have an opportunity to submit a response to reviewers and update the papers during the rebuttal period. Reviews and rebuttals of accepted papers will be made publicly available. During the review/rebuttal phase all papers are private and anonymized; rejected papers will remain private after decisions are announced.
We take a broad view of robot learning. Papers with both experimental and theoretical results relevant to robot learning are welcome. Our intent is to keep CoRL a selective top-tier conference on robotic learning.
All submissions are confidential. Do not discuss, share, or store assigned papers outside the OpenReview system, and do not use them for any purpose other than reviewing. If you identify any conflict of interest with a paper assigned to you (institutional, advisor/advisee, recent collaborator, personal), notify your Area Chair immediately so the paper can be reassigned.
Process: ACs will identify the desk rejection candidates, using one of the criteria below as a justification. SACs will examine the candidates and make a final proposition to the PCs. We will err on the side of caution, and only desk reject papers when there is a consensus between all PCs and the SACs.
The paper can be desk rejected for one of the following reasons: not meeting the reciprocal reviewing requirement, formatting issues, anonymity violation, or scope.
Reciprocal reviewing requirement: please check the Instructions for Authors section for this requirement.
Formatting issues: The paper is either too long or in an incorrect format.
Anonymity violation: The main manuscript, supplemental materials, or a link provided in a paper identifies one or more of the authors.
Scope: All CoRL submissions must demonstrate the relevance to Robot Learning through
Intent: Explicitly address a learning question for physical robots OR
Outcome: Evaluate the proposed learning solution. This should mainly be done on physical robots, though realistic simulation might also be acceptable.
Rejection examples
No learning: Manually design and tune the performance of a robot controller without use of learning.
No learning: A search algorithm for model-based planning.
No robotics: A generic result on sample complexity.
No robotics: A generic RL algorithm.
Little robotics: Improved performance on a standard CV dataset, e.g., ImageNet recognition.
Gray
An RL algorithm that works only in simulator X. Does it transfer to real robot learning (sim2real, data efficiency, …)? Yes for CARLA for autonomous driving. No for Cheetah/Human-oid in Mujuco. According to our stated principles, the submission satisfies the intent. Its failure or success to demonstrate the relevance will be determined during the review process.
New this Year: First-round rejection
Papers that do not receive at least one score of weak accept or above from either a reviewer or the Area Chair will be candidates for first-round rejection without rebuttal. Senior Area Chairs and Program Chairs will review all such candidates before any decision is finalized.
Rationale. When all reviewers and the AC agree a paper falls below the acceptance threshold, a rebuttal rarely changes the outcome. This policy focuses the rebuttal and discussion period on papers where substantive engagement can make a difference, and gives authors an earlier signal to revise and resubmit in another venue.
Implication for reviewers. Your initial scores now carry greater weight. Before assigning a score below weak accept, ask yourself: could the issues you identify reasonably be addressed in a rebuttal, or do they reflect deeper shortcomings? Score honestly and do not inflate, but be deliberate, and ensure your review text is thorough enough to stand on its own, as the authors may not have the opportunity to respond.
Reviewers should have expertise in robotics, robot learning, machine learning, or complementary fields, as demonstrated by a record of previous publications in the area. Reviewers ideally have a minimum of three first-author publications in major robotics conferences and journals (e.g., CoRL, RSS, ICRA, IROS, IJRR, T-RO, RA-L) and core machine learning venues (e.g., NeurIPS, ICML, AISTATS, JMLR, PAMI). To recruit more experienced reviewers, Area Chairs were asked to nominate faculty members, research scientists, researchers, and senior graduate students.
The Use of AI Tools (New this year)
Reviewers may use AI tools to support their reviewing work, but the review must reflect the reviewer’s own informed judgment. We expect reviewers to first read each paper and form an initial assessment independently. AI assistance is permitted for subsequent tasks such as brainstorming, exploring related literature, clarifying unfamiliar technical concepts, or refining the wording of the written review. Even when AI-assisted exploration reshapes a reviewer’s view of the work, the final assessment and recommendation must be that of the reviewer, and the reviewer bears ultimate responsibility for the accuracy, rigor, and quality of the review submitted. To preserve confidentiality, reviewers must not upload the manuscript or any portion of it to AI services that do not provide enterprise-level confidentiality guarantees, as submissions are privileged material entrusted to reviewers under the CoRL peer-review terms.
Reviewers who use AI tools at any stage of the reviewing process must disclose this to the Program Committee, indicating which tools were used and for what purpose (e.g., literature search, language polishing, conceptual clarification).
Similarly, authors are required to disclose any significant usage of AI tools in research ideation or writing. If undisclosed AI-tool usage is uncovered during discussion, please notify the Area Chair.
After receiving the reviews, authors of the papers are invited to submit a 1-page rebuttal in pdf, by the resubmission deadline. The rebuttal should be focused on addressing core concerns identified by the reviewers and the AC. All the rebuttals will be reviewed by the reviewer, AC and SAC. All papers will be discussed in dedicated SAC-ACs meetings, as well as a final SACs-PCs meeting that will finalize the decisions.
New this year: post-rebuttal box. Reviewers must fill out a post-rebuttal field in the review form. State whether the rebuttal addressed your concerns, update your score if it has changed, and justify your final recommendation. This field is required and is the main input the AC uses during discussion.
After receiving the reviews, authors of the papers are invited to submit a 1-page rebuttal in pdf, by the resubmission deadline. The rebuttal should be focused on addressing any factual errors in the review instead of providing additional experimental results. Reviewers should not request additional experimental results in rebuttalTimelAll the rebuttals will be reviewed by the reviewer, original AC, and potentially by external reviewers. The AC reserves the right to invite new reviewers if needed. The results of the current review(s) will be shared with the new reviewers in such cases
Paper bidding (optional): May 30 – June 9, 2026
Reviewer assignments released by: June 12, 2026
Reviewing period (firm deadline): June 15 – July 13, 2026
Reviews released to authors: August 4, 2026
Author rebuttal period: August 4 – August 11, 2026
Reviewer ↔ AC discussion: August 12 – August 19, 2026
Post-rebuttal review update deadline: August 19, 2026
Paper acceptance notification: September 4, 2026
The July 13 reviewing deadline is firm. Late reviews place a disproportionate burden on ACs and SACs and may trigger emergency reviewer reassignment. If you foresee a conflict (illness, unavoidable travel, late-discovered COI), contact your AC as early as possible — advance notice is much easier to handle than a missing review at the deadline.
During the reviewer–AC discussion window (August 12 – 19), please remain responsive to AC messages and update your review and score if the rebuttal warrants it. We require you to explicitly acknowledge that you read the rebuttal and you reply at the respective rebuttal box.