# reviewpaper.ai > Simulated peer reviews of your paper in your target venue's exact review form, before you submit. Upload a paper PDF and pick a venue (NeurIPS, ICLR, ICML, CVPR, ACL Rolling Review, AAAI, KDD, TMLR and 20 more). Three simulated reviewers with different personas fill in that venue's reviewer form (scores, strengths, weaknesses, questions), an area chair writes a meta-review, and you get an expected score range, an accept / borderline / reject outlook, a ranked "fix before submitting" list and the venue checklist. Every review point links to the passage in the PDF it refers to. The whole paper is read, appendices included. You can then chat with the paper (answers cite the page). - Sample report (no sign-up): https://reviewpaper.ai/sample - Pricing: https://reviewpaper.ai/pricing.md - Review forms for 29 venues: https://reviewpaper.ai/venues.md - AI conference deadlines: https://reviewpaper.ai/deadlines.md - Guides: https://reviewpaper.ai/guides - Free tools: https://reviewpaper.ai/tools --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # reviewpaper.ai pricing - 3 free credits at sign-up, no card needed. - Pro: $10/month for 12 credits. - Credit pack: $5 for 5 credits. Credits per review run, by reviewer model: | Tier | Model | Credits | |---|---|---| | Standard | Claude Sonnet 5.5 | 1 | | Deep | Claude Opus 5.5 | 2 | | Max | Claude Fable 5.1 | 6 | Web page: https://reviewpaper.ai/pricing --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # NeurIPS 2026 review form > Simulate NeurIPS 2026 main-track reviews: Summary, Strengths and Weaknesses, Quality / Clarity / Significance / Originality (1–4), the 1–6 overall score and confidence, plus a paper-checklist audit. - Venue: Conference on Neural Information Processing Systems - Form status: official (from the venue's public reviewer form) - Official source: https://neurips.cc/Conferences/2026/MainTrackHandbook - Overall score: Overall, 1–6 - Rough acceptance bar (our simulation): mean ≥ 4 - Web page: https://reviewpaper.ai/venues/neurips ## How reviewers read the paper NeurIPS asks for a narrative assessment framed as reasons to accept or reject along four criteria (Quality, Clarity, Significance, Originality), each also scored 1–4, then 3–5 actionable questions that say what would change the score, a limitations check that does not penalize candor, a 1–6 overall score (borderline 3/4 used sparingly) and a 1–5 confidence. In 2026 reviewers also confirm the paper's declared contribution type (General, Theory, Use-Inspired, Concept & Feasibility, Negative Results), which changes how the criteria are weighted. - Write narrative prose: each strength/weakness is one or more full, substantiated sentences, not a bullet fragment. A good review is more than simply bullet points. - Fair, precise and specific; superficial claims without evidence are worse than no review. - Never rude. Frame criticism constructively. - Do not lower scores for formatting issues; report them in the formatting field. **Where the bar sits.** NeurIPS 2025 accepted 24.5% of main-track papers. Accepted papers averaged about 4.3 and rejected about 3.6 (Paper Copilot). Rough rule: mean ≥ 4.0 likely accept, 3.5–4.0 borderline, < 3.5 likely reject. ## Reviewer form, field by field ### 1. Summary Type: free text. Briefly summarize the paper and its contributions in your own words. Do not critique here and do not paste the abstract. The authors should agree with a well-written summary. ### 2. Contribution Type Type: single choice. Select the contribution type the paper most plausibly declared. The type changes how the criteria are read: Theory → correctness of claims is primary and empirical validation is not required; Use-Inspired → non-standard datasets are fine; Concept & Feasibility → high bar on significance and originality; Negative Results → high bar, the result must be surprising. - General - Theory - Use-Inspired - Concept & Feasibility - Negative Results ### 3. Strengths Type: list of points. Part of "Strengths and Weaknesses". Reasons to accept, across Quality (technical soundness, claims supported by theory or experiments), Clarity (clear writing, enough detail to reproduce), Significance (impact, likelihood of follow-up), and Originality (new insights, clear differentiation from prior work; novel evaluation or efficiency gains also count). Write substantiated narrative sentences, not bare bullet fragments. ### 4. Weaknesses Type: list of points. Part of "Strengths and Weaknesses". Reasons to reject across the same four dimensions. Be specific and substantiate each point with evidence from the paper. Distinguish fatal flaws from fixable issues. ### 5. Quality Type: score, 1–4. Technical soundness; claims supported; appropriate methods; honesty about weaknesses. - **4**: excellent - **3**: good - **2**: fair - **1**: poor ### 6. Clarity Type: score, 1–4. Writing and organization; enough detail for an expert to reproduce. - **4**: excellent - **3**: good - **2**: fair - **1**: poor ### 7. Significance Type: score, 1–4. Impact on the community; likelihood of use or follow-up. - **4**: excellent - **3**: good - **2**: fair - **1**: poor ### 8. Originality Type: score, 1–4. New insights, methods, tasks or combinations; clear differentiation from prior work. - **4**: excellent - **3**: good - **2**: fair - **1**: poor ### 9. Questions Type: list of points. Ideally 3–5 actionable questions and suggestions for the authors. For each, state what answer would change your score up or down. ### 10. Limitations Type: free text. Have the authors adequately addressed the limitations and potential negative societal impact of their work? If yes, write "yes". Otherwise give constructive suggestions. Do not penalize candor about limitations. ### 11. Overall Type: score, 1–6. Overall recommendation. Borderline scores (3, 4) should be used sparingly. - **6**: Strong Accept: Technically flawless paper with groundbreaking impact on at least one area of AI. - **5**: Accept: Technically solid paper, with high impact on at least one sub-area of AI or moderate-to-high impact on more than one area. - **4**: Borderline accept: Technically solid paper where reasons to accept outweigh reasons to reject, e.g., limited evaluation. Please use sparingly. - **3**: Borderline reject: Technically solid paper where reasons to reject, e.g., limited evaluation, outweigh reasons to accept. Please use sparingly. - **2**: Reject: For instance, a paper with technical flaws, weak evaluation, inadequate reproducibility and incompletely addressed ethical considerations. - **1**: Strong Reject: For instance, a paper with well-known results or unaddressed ethical considerations. ### 12. Confidence Type: score, 1–5. How certain you are about your assessment. - **5**: You are absolutely certain about your assessment. You are very familiar with the related work and checked the math/other details carefully. - **4**: You are confident in your assessment, but not absolutely certain. It is unlikely, but not impossible, that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. - **3**: You are fairly confident in your assessment. It is possible that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked. - **2**: You are willing to defend your assessment, but it is quite likely that you did not understand the central parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked. - **1**: Your assessment is an educated guess. The submission is not in your area or the submission was difficult to understand. Math/other details were not carefully checked. ### 13. Ethical Concerns Type: single choice. Choose exactly one option. A clear concern must also be described under Weaknesses. - NO or VERY MINOR ethics concerns only - CLEAR MAJOR CONCERN - Major Concern: Improper research involving human subjects - Major Concern: Data privacy, copyright, and consent - Major Concern: Data quality and representativeness - Major Concern: Safety and security - Major Concern: Discrimination, bias, and unfairness - Major Concern: Deception and harassment - Major Concern: Environmental impact - Major Concern: Human rights (including surveillance) ### 14. Paper Formatting Concerns Type: free text. Report major formatting violations (page limit, anonymity, font). Write "None" if there are none. Do not lower the overall score for them. ## Area chair meta-review NeurIPS ACs triage papers into heading-to-reject / needs clarification / definite accept, then write a meta-review that explicitly explains the decision. Decide from arguments, not the average score. Recommendation options: Accept (oral) · Accept (spotlight) · Accept (poster) · Reject ## NeurIPS Paper Checklist 1. Claims: do the abstract and introduction accurately reflect the paper's contributions and scope? 2. Limitations: does the paper discuss the limitations of the work? 3. Theory, Assumptions and Proofs: are full assumptions and complete, correct proofs provided for each theoretical result? 4. Experimental Result Reproducibility: is enough information disclosed to reproduce the main results? 5. Open Access to Data and Code: are data and code provided with instructions to reproduce? 6. Experimental Setting/Details: are training and test details (splits, hyperparameters, how they were chosen, optimizer) specified? 7. Experiment Statistical Significance: are error bars or other statistical-significance information reported and correctly defined? 8. Experiments Compute Resources: is the compute (hardware, time, memory) needed to reproduce each experiment reported? 9. Code of Ethics: does the research conform to the NeurIPS Code of Ethics? 10. Broader Impacts: are potential positive and negative societal impacts discussed? 11. Safeguards: are safeguards described for responsible release of high-risk data or models? 12. Licenses (for existing assets): are creators of existing assets credited and licenses respected? 13. Assets (new assets): are new assets well documented, with documentation provided alongside them? 14. Crowdsourcing and Research with Human Subjects: are instructions, screenshots and compensation details included? 15. IRB Approvals: are risks to participants described and IRB (or equivalent) approval stated? 16. Declaration of LLM usage: is any important, original or non-standard use of LLMs in the core methods described? ## LLM policy NeurIPS 2026 runs a sanctioned-LLM experiment: on some assigned papers reviewers may use only the sanctioned LLM; on all others any LLM use is forbidden, and sharing a submission with an unsanctioned LLM breaks confidentiality. Official policy: https://neurips.cc/Conferences/2026/MainTrackHandbook --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=neurips (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # ICLR 2026 review form > Simulate ICLR 2026-style OpenReview reviews: Summary, Soundness / Presentation / Contribution (1–4), Strengths, Weaknesses, Questions, the new even-numbered 0–10 rating and confidence. - Venue: International Conference on Learning Representations - Form status: modeled, not official. ICLR's guide is public, but the OpenReview form itself is not. Field order follows the long-standing ICLR form and the 2026 rating labels (0/2/4/6/8/10) come from secondary sources. - Official source: https://iclr.cc/Conferences/2026/ReviewerGuide - Overall score: Rating, 0–10 - Rough acceptance bar (our simulation): mean ≥ 5 - Web page: https://reviewpaper.ai/venues/iclr ## How reviewers read the paper ICLR's reviewer guide prescribes: summarize the claimed contribution; list strong and weak points; clearly state an initial recommendation with one or two key reasons; give supporting arguments; ask questions; optionally add feedback marked as not part of the decision. Reviewers should ask what problem is tackled, whether the approach is well motivated and placed in the literature, whether claims are supported correctly and rigorously, and whether the work is significant. Reviews are public on OpenReview. - Constructive tone — criticize with kindness; reviews are public. - Make the initial recommendation and its one or two key reasons explicit at the top of the weaknesses. - As comprehensive as possible; excessively short reviews get flagged. - State-of-the-art results are not required for a contribution. **Where the bar sits.** ICLR 2026 accepted 27.4% of decided submissions. Paper Copilot means: posters ≈ 5.4, orals ≈ 6.4, rejected ≈ 4.0. Rough rule: mean ≥ 5 likely accept, ~4.4–5 borderline. ## Reviewer form, field by field ### 1. Summary Type: free text. Summarize what the paper claims to contribute. Be positive and constructive. ### 2. Soundness Type: score, 1–4. Are the claims adequately supported, and are the results technically correct and rigorous? - **4**: excellent - **3**: good - **2**: fair - **1**: poor ### 3. Presentation Type: score, 1–4. How clearly the paper communicates its ideas, methods and results. - **4**: excellent - **3**: good - **2**: fair - **1**: poor ### 4. Contribution Type: score, 1–4. Does it provide new, relevant, valuable knowledge to the ICLR community? State-of-the-art results are not required. - **4**: excellent - **3**: good - **2**: fair - **1**: poor ### 5. Strengths Type: list of points. Strong points, as comprehensive as possible, each substantiated. ### 6. Weaknesses Type: list of points. Weak points, as comprehensive as possible. Lead with the one or two key reasons behind your recommendation, then supporting arguments. ### 7. Questions Type: list of points. Questions that would clarify your understanding or supply missing evidence. Additional feedback not part of the decision may be marked as such. ### 8. Flag For Ethics Review Type: single choice. Choose one. - No ethics review needed. - Yes, Discrimination / bias / fairness concerns - Yes, Privacy, security and safety - Yes, Legal compliance (e.g., GDPR, copyright, terms of use) - Yes, Potentially harmful insights, methodologies and applications - Yes, Responsible research practice (e.g., human subjects, data release) - Yes, Research integrity issues (e.g., plagiarism, dual submission) - Yes, Other reasons ### 9. Rating Type: score, 0–10 (step 2). Initial recommendation. Clearly state accept or reject with one or two key reasons in the review text. - **10**: strong accept, should be highlighted at the conference - **8**: accept, good paper - **6**: marginally above the acceptance threshold - **4**: marginally below the acceptance threshold - **2**: reject, not good enough - **0**: strong reject ### 10. Confidence Type: score, 1–5. How certain you are about your assessment. - **5**: You are absolutely certain about your assessment. You are very familiar with the related work and checked the math/other details carefully. - **4**: You are confident in your assessment, but not absolutely certain. It is unlikely, but not impossible, that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. - **3**: You are fairly confident in your assessment. It is possible that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked. - **2**: You are willing to defend your assessment, but it is quite likely that you did not understand the central parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked. - **1**: You are unable to assess this paper and have alerted the ACs to seek an opinion from different reviewers. ## Area chair meta-review ICLR meta-reviews are public. Base the decision on the arguments in the reviews, not the mean score. Recommendation options: Accept (Oral) · Accept (Poster) · Reject ## ICLR author statements 1. Ethics statement (recommended, max 1 page, outside the page limit) 2. Reproducibility statement pointing to code, proofs and experimental details (strongly encouraged) 3. LLM usage disclosure if LLMs played a significant role in ideation or writing (required; non-disclosure can lead to desk rejection) 4. Main text within 9 pages at submission (references and appendix unlimited) 5. Anonymized submission (no identifying links or acknowledgements) ## LLM policy ICLR 2026 allows LLMs as general-purpose writing assistance, but reviewers are fully responsible for content and must disclose LLM use in the review form. Reviews were screened with LLM-text detectors. Official policy: https://iclr.cc/Conferences/2026/ReviewerGuide --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=iclr (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # ICML 2026 review form > Simulate ICML 2026 main-track reviews: Summary, Strengths and Weaknesses, Soundness / Presentation / Significance / Originality (1–4), Key Questions For Authors, the 1–6 Overall Recommendation and confidence. - Venue: International Conference on Machine Learning - Form status: official (from the venue's public reviewer form) - Official source: https://icml.cc/Conferences/2026/ReviewerInstructions - Overall score: Overall Recommendation, 1–6 - Rough acceptance bar (our simulation): mean ≥ 4 - Web page: https://reviewpaper.ai/venues/icml ## How reviewers read the paper ICML 2026 reviewers summarize the paper without critique, write a substantiated Strengths and Weaknesses assessment covering soundness, presentation, significance and originality (each also scored 1–4, with fair/poor ratings justified), ask 3–5 numbered key questions that say how each answer would change the score, check limitations, and give a 1–6 overall recommendation (3 and 4 sparingly) plus confidence. - Constructive, specific and substantiated. - Number the key questions and, for each, state how possible answers would move your score. - Every fair/poor sub-score must be justified in the weaknesses. **Where the bar sits.** ICML 2026 accepted about 26.6% of submissions. No public score statistics; inferred from NeurIPS's identical scale: mean ≥ 4 likely accept, 3.5–4 borderline. ## Reviewer form, field by field ### 1. Summary Type: free text. Briefly summarize the paper and its contributions in your own words. No critique, no pasted abstract. ### 2. Strengths Type: list of points. Part of "Strengths and Weaknesses". Cover soundness, presentation, significance and originality. Be specific and substantiate every point. ### 3. Weaknesses Type: list of points. Part of "Strengths and Weaknesses". Be specific and substantiate every point. Any Soundness/Presentation/Significance/Originality rating of fair or poor must be justified here. ### 4. Soundness Type: score, 1–4. Technical claims, methodology, and whether the central claims are supported with evidence. - **4**: excellent - **3**: good - **2**: fair - **1**: poor ### 5. Presentation Type: score, 1–4. Clarity, structure, and positioning relative to prior work. - **4**: excellent - **3**: good - **2**: fair - **1**: poor ### 6. Significance Type: score, 1–4. Importance of the problem and likely influence. - **4**: excellent - **3**: good - **2**: fair - **1**: poor ### 7. Originality Type: score, 1–4. New insights, methods, theory, data or perspectives; novel combinations and evaluations count. - **4**: excellent - **3**: good - **2**: fair - **1**: poor ### 8. Key Questions For Authors Type: list of points. Ideally 3–5 numbered questions whose answers would change your evaluation. For each, say how the possible answers would change your score. ### 9. Limitations Type: free text. Write "yes" if limitations and potential negative societal impact are adequately discussed; otherwise constructive suggestions. ### 10. Overall Recommendation Type: score, 1–6. Overall recommendation. 3 and 4 should be used sparingly. - **6**: Strong Accept: Technically flawless paper with exceptional impact on one or more areas of AI, with strong evaluation, reproducibility, and resources, and no unaddressed ethical considerations. - **5**: Accept: Technically solid paper, with high impact on at least one sub-area of AI or moderate-to-high impact on more than one area of AI, with good-to-excellent evaluation, resources, reproducibility, and no unaddressed ethical considerations. - **4**: Weak accept: Technically solid paper that advances at least one sub-area of AI, with a contribution that others are likely to build on, but with some weaknesses that limit its impact (e.g., limited evaluation). Please use sparingly. - **3**: Weak reject: A paper with clear merits, but also some weaknesses, which overall outweigh the merits. Papers in this category require revisions before they can be meaningfully built upon by others. Please use sparingly. - **2**: Reject: For instance, a paper with technical flaws, weak evaluation, inadequate reproducibility, incompletely addressed ethical considerations, or writing so poor that it is not possible to understand its key claims. - **1**: Strong Reject: For instance, a paper with well-known results, unaddressed ethical considerations, or a poorly written paper where it is impossible to tell what the nature of its contribution is. ### 11. Confidence Type: score, 1–5. How certain you are about your assessment. - **5**: You are absolutely certain about your assessment. You are very familiar with the related work and checked the math/other details carefully. - **4**: You are confident in your assessment, but not absolutely certain. It is unlikely, but not impossible, that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. - **3**: You are fairly confident in your assessment. It is possible that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked. - **2**: You are willing to defend your assessment, but it is quite likely that you did not understand the central parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked. - **1**: Your assessment is an educated guess. The submission is not in your area or the submission was difficult to understand. Math/other details were not carefully checked. ### 12. Ethical Concerns Type: free text. "No" if none. Otherwise name the area (discrimination/bias/fairness; inappropriate applications and impact; responsible research practice; privacy and security; legal compliance; research integrity; other) and explain. ## Area chair meta-review ICML ACs decide from concrete arguments, not average scores, and use a fixed template when rejecting. Recommendation options: Accept (oral) · Accept (spotlight poster) · Accept (poster) · Reject ## ICML submission requirements 1. Impact Statement included (main track; does not count toward the page limit) 2. Main text within 8 pages at submission 3. Anonymized submission 4. Reproducibility details: hyperparameters, seeds, compute, and code availability 5. Limitations discussed ## LLM policy ICML 2026 assigns each reviewer Policy A (no LLMs) or Policy B (LLMs may help understand the paper and polish wording, but may not identify strengths/weaknesses or write the review). Violations can get the reviewer's own papers desk-rejected. Official policy: https://icml.cc/Conferences/2026/LLM-Policy --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=icml (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # AISTATS 2026 review form > Simulate AISTATS 2026 reviews: Summary and Contributions, Soundness / Significance / Novelty / Prior Work / Reproducibility choices with justifications, Clarity, the 1–7 rating and 1–5 confidence. - Venue: International Conference on Artificial Intelligence and Statistics - Form status: official (from the venue's public reviewer form) - Official source: https://virtual.aistats.org/Conferences/2026/ReviewerGuidelines - Overall score: Rating, 1–7 - Rough acceptance bar (our simulation): mean ≥ 5 - Web page: https://reviewpaper.ai/venues/aistats ## How reviewers read the paper AISTATS 2026 reviewers summarize the paper, then choose Soundness, Significance and Novelty options each with a written justification (incremental or known-results verdicts must cite the prior work), note non-conventional contributions, assess clarity, relation to prior work and reproducibility, and give a 1–7 rating and 1–5 confidence. Reviewers should be thoughtful, fair, specific and timely, avoid dismissive criticism of novel or cross-disciplinary work, and respond to every rebuttal. - Thoughtful, fair, useful and specific; avoid dismissive criticism, especially of novel or cross-disciplinary work. - Cite the specific prior publications whenever you call a result incremental or known. - Justify every Soundness / Significance / Novelty choice. **Where the bar sits.** The AISTATS 2026 call says acceptance tends to be around 25% (2025: 31%, 2024: 28%). No score statistics are public; on the 1–7 scale a mean of 5 (Borderline Accept) or more is a rough proxy for acceptance, 4–5 borderline. ## Reviewer form, field by field ### 1. Summary And Contributions Type: free text. Briefly summarize the paper and its contributions. ### 2. Soundness Type: single choice. Choose one; justify it in the strengths/weaknesses. - Correct / minor errors (e.g., typo or errors that do not affect the main results) - Major errors (e.g., an incorrect theorem or derivation) - Not applicable ### 3. Significance Type: single choice. Choose one; justify it in the strengths/weaknesses. - Significant contributions - Somewhat significant - Not significant - Not applicable ### 4. Novelty Type: single choice. Choose one. "Incremental" or "Known results" requires citing the prior publications. - Ground-breaking - New results - Incremental compared to existing results - Known results, or a trivial extension of known results ### 5. Justification Of Soundness, Significance And Novelty Type: list of points. The form has a justification box under each of Soundness, Significance and Novelty. Give the specific evidence behind each choice: errors found (with location), why the contribution matters or does not, and the prior work it builds on (cited). ### 6. Non-conventional Contributions Type: free text. Contributions outside the usual mould, e.g. bridging disciplines or posing new problems. "None" if not applicable. ### 7. Clarity Type: free text. Is the paper clearly written? Consider whether it (1) clearly states its contributions, notation and results, (2) supports the claims made in the title, abstract and introduction with results, (3) explains the meaning of its theoretical assumptions, and/or (4) motivates the proposed methodology well, e.g. with examples. ### 8. Relation To Prior Work Type: single choice. Choose one. Name any missing work under Additional Comments. - All related works are clearly discussed - All related works are cited but the discussion is less clear - Some minor related works are missing or described incorrectly - Some important related works are missing or described incorrectly ### 9. Additional Comments Type: list of points. Feedback and questions for the authors, including any missing related work. ### 10. Reproducibility Type: single choice. Choose one. - Sufficient amount of details available for reproducing the main results - Some amount of details available - Insufficient amount of details available - Not applicable (the paper does not have experiments) ### 11. Rating Type: score, 1–7. Overall rating. Note the official order: 4 ("Weak Accept / Weak Reject") sits between Borderline Reject and Borderline Accept; use it sparingly. - **7**: Strong Accept (Technically excellent paper with clear novelty and broad potential impact beyond a single sub-area of AI and stats. The work is very well executed with convincing evaluation and good attention to reproducibility and ethical considerations.) - **6**: Accept (Technically solid paper, with high impact on at least one sub-area. The contribution is convincing and the evaluation is adequate, though there may be some limitations or open questions.) - **5**: Borderline Accept (Technically solid paper where reasons to accept outweigh reasons to reject, e.g., limited evaluation.) - **4**: Weak Accept / Weak Reject (Paper has some merits but also notable weaknesses. Reasons to accept and reject are roughly balanced, and the paper may require further evaluation or discussion. Please use sparingly.) - **3**: Borderline Reject (Technically solid paper where reasons to reject, e.g., limited evaluation, outweigh reasons to accept, e.g., good evaluation.) - **2**: Reject (For instance, a paper with technical flaws, weak evaluation, inadequate reproducibility and incompletely addressed ethical considerations.) - **1**: Strong Reject (For instance, a paper with well-known results or unaddressed ethical considerations.) ### 12. Confidence Type: score, 1–5. How certain you are about your assessment. - **5**: Absolutely certain (Use this sparingly; please ensure you are very familiar with the related work, you have carefully checked and understood the proofs and/or experimental details if applicable) - **4**: Confident, but not absolutely certain. (It is unlikely, but not impossible, that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work.) - **3**: Fairly confident (It is possible that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked.) - **2**: Somewhat confident (You are willing to defend your assessment, but it is quite likely that you did not understand central parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked.) - **1**: Educated guess (The submission is not in your area or the submission was difficult to understand. Math/other details were not carefully checked.) ### 13. Ethical Concerns Type: single choice. Choose one (visible to PC/SAC/AC only). - No concern - Minor concerns - Major concerns ## Area chair meta-review AISTATS ACs recommend accept/reject and nominate orals. The meta-review form itself is not public. Recommendation options: Accept (oral) · Accept (poster) · Reject ## AISTATS reproducibility checklist (in the paper, after the references) 1. For all models and algorithms: mathematical setting, assumptions and algorithm described; properties and complexity (time, space, sample size) analysed; anonymized source code (optional) 2. For theoretical results: all assumptions stated; complete proofs; clear explanations of any assumptions 3. For experiments: code, data and instructions to reproduce; training details (splits, hyperparameters, how chosen); error bars or variance over runs; compute resources 4. For existing assets used: creators cited; license stated; new assets documented and released 5. For crowdsourcing or human-subjects research: full instructions; potential risks and IRB approval; compensation ## LLM policy AISTATS 2026 strictly prohibits generative AI, including LLMs, in reviewing: reviewers may not use it to write reviews or put any submission or review content into it. Official policy: https://virtual.aistats.org/Conferences/2026/ReviewerGuidelines --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=aistats (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # UAI 2026 review form > Simulate UAI 2026 reviews: Summary and contributions, five 1–4 assessments (originality, correctness, evidence, reproducibility, clarity), main strengths and weaknesses, detailed comments, and the 1–10 overall score with justification and confidence. - Venue: Conference on Uncertainty in Artificial Intelligence - Form status: official (from the venue's public reviewer form) - Official source: https://auai.org/uai2026/instructions_for_reviewers_and_area_chairs - Overall score: Overall Score, 1–10 - Rough acceptance bar (our simulation): mean ≥ 6 - Web page: https://reviewpaper.ai/venues/uai ## How reviewers read the paper UAI 2026 reviewers summarize the motivation and contributions briefly (mainly for ACs), rate five axes on 1–4 (originality, correctness, support of claims by evidence, reproducibility, clarity), list main strengths and weaknesses along those axes, give detailed comments, and choose a 1–10 overall score with a short justification and a 1–5 confidence. Judge the scientific contribution, not personal interest in the topic. Ethics flags are separate from the score. - Detailed, specific and polite; avoid vague or subjective complaints. - Judge the scientific contribution, not whether the topic interests you. - Respect the character limits: summary and justification are a few lines each. - Use the borderline scores 4 and 5 sparingly. **Where the bar sits.** UAI historically accepts about 30% of submissions; no 2026 statistics are public. The reviewer instructions target at most 25% of papers at 7 (Accept) or above, so a mean around 6 is competitive and 5–6 is borderline. ## Reviewer form, field by field ### 1. Summary And Contributions Type: free text. At most 500 characters: the paper's motivation and key contributions in a few lines. Contributions may be theoretical, methodological, algorithmic, applied, empirical, connecting ideas across fields, or a critical analysis. ### 2. Originality/Novelty Type: score, 1–4. Part of "Assessment Of The Paper". How new are the main ideas? - **4**: Excellent: The main ideas of the paper are ground-breaking. - **3**: Good: The paper makes non-trivial advances over the current state-of-the-art. - **2**: Fair: The paper contributes some new ideas. - **1**: Poor: The main ideas of the paper are not novel or represent incremental progress. ### 3. Correctness/Technical Quality Type: score, 1–4. Part of "Assessment Of The Paper". Technical soundness of the methods, proofs and experiments. - **4**: Excellent: I am confident that the paper is technically sound, and I have carefully checked the details. - **3**: Good: The paper appears to be technically sound, but I have not carefully checked the details. - **2**: Fair: The paper has minor, easily fixable, technical flaws that do not impact the validity of the main results. - **1**: Poor: The paper has major technical flaws. ### 4. Extent To Which Claims Are Supported By Evidence Type: score, 1–4. Part of "Assessment Of The Paper". Do experiments, proofs, code and references back up the claims? - **4**: Excellent: all claims are supported by very convincing evidence (in the form of comprehensive experimental evaluation, rigorous mathematical proofs, detailed (pseudo-)code, precise references, well-motivated and realistic assumptions) and the authors deliver what they promise. - **3**: Good: the main claims are supported by convincing evidence (in the form of adequate experimental evaluation, proofs, (pseudo-)code, references, assumptions). - **2**: Fair: the main claims are somewhat supported by evidence (but the experimental evaluation may be weak, or does not match entirely with the claims, important baselines may be missing, proofs contain important ideas but lack rigor, algorithmic details are only discussed superficially, references are imprecise, assumptions are not sufficiently motivated or explicated, etc.). - **1**: Poor: the authors fail to convincingly back up their main claims (e.g., if the experimental evaluation is flawed, proofs are lacking or invalid, references are missing, assumptions are not realistic, not specified, or not motivated). ### 5. Reproducibility Type: score, 1–4. Part of "Assessment Of The Paper". Could competent researchers reproduce the main results? - **4**: Excellent: key resources (e.g. proofs, code, data) are available and key details (e.g. proof sketches, experimental setup) are comprehensively described for competent researchers to confidently and easily reproduce the main results. - **3**: Good: key resources (e.g. proofs, code, data) are available and key details (e.g. proofs, experimental setup) are sufficiently well-described for competent researchers to confidently reproduce the main results. - **2**: Fair: key resources (e.g. proofs, code, data) are unavailable but key details (e.g. proof sketches, experimental setup) are sufficiently well-described for an expert to confidently reproduce the main results. - **1**: Poor: key details (e.g. proof sketches, experimental setup) are incomplete/unclear, or key resources (e.g. proofs, code, data) are unavailable. ### 6. Clarity Of Writing Type: score, 1–4. Part of "Assessment Of The Paper". Organization and clarity of the writing. - **4**: Excellent: The paper is well-organized and clearly written. - **3**: Good: The paper is well organized but the presentation could be improved. - **2**: Fair: The paper is somewhat clear, but some important details are missing or unclear. - **1**: Poor: The paper is unclear and very hard to understand. ### 7. Main Strengths Type: list of points. At most 3000 characters, organized along the five assessment axes above. ### 8. Main Weaknesses Type: list of points. At most 3000 characters, organized along the five assessment axes above. Be specific; avoid vague or subjective complaints. ### 9. Detailed Comments To The Authors Type: list of points. Detailed, specific and polite comments, questions and suggestions for improving the paper. ### 10. Overall Score Type: score, 1–10. Ideally no more than 25% of papers at Accept (7) or above, 20% at Strong Accept or above, 10% at Very Strong Accept or above and 1% at Award quality. 4 and 5 sparingly. - **10**: Award quality - **9**: Very Strong Accept - **8**: Strong Accept - **7**: Accept - **6**: Weak Accept - **5**: Borderline accept. Please use sparingly. - **4**: Borderline reject. Please use sparingly. - **3**: Reject - **2**: Strong Reject - **1**: Very Strong Reject ### 11. Justification For Your Score Type: free text. At most 500 characters: the decisive reasons for the overall score. ### 12. Confidence In Your Score Type: score, 1–5. How confident you are in your score. - **5**: Very confident - **4**: Quite confident - **3**: Somewhat confident - **2**: Not very confident - **1**: Not confident ### 13. Ethical Concerns (optional) Type: free text. "No", or "Yes" with an explanation. The overall score should not depend on this answer. ## Area chair meta-review UAI ACs run the discussion, write the meta-review and recommend accept or reject to the PCs. The meta-review form itself is not public. Recommendation options: Accept · Reject ## LLM policy UAI 2026 does not allow LLMs to write reviews or meta-reviews; they may be used only to check grammar, and the reviewer is responsible for all submitted text. Official policy: https://auai.org/uai2026/instructions_for_reviewers_and_area_chairs --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=uai (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # TMLR 2026 review form > Simulate TMLR reviews: Summary of contributions, the two acceptance questions (are the claims supported by evidence? would some of TMLR's audience care?), requested changes marked critical or optional, broader-impact concerns, and the Accept / Leaning Accept / Leaning Reject / Reject recommendation. - Venue: Transactions on Machine Learning Research - Form status: official (from the venue's public reviewer form) - Official source: https://jmlr.org/tmlr/reviewer-guide.html - Overall score: Recommendation, 1–4 - Rough acceptance bar (our simulation): mean ≥ 3 - Web page: https://reviewpaper.ai/venues/tmlr ## How reviewers read the paper TMLR is a journal with rolling submissions. Acceptance rests on two questions only: are the claims supported by accurate, convincing and clear evidence, and would at least some of TMLR's audience be interested in the findings? TMLR explicitly avoids judging novelty, significance or impact: a narrow but well-supported contribution should be accepted. Presentation problems matter only if they materially get in the reader's way. Reviewers list requested changes, each marked critical or merely strengthening, flag broader-impact concerns, and after discussion recommend Accept / Leaning Accept / Leaning Reject / Reject (plus optional certifications). Reviews are public. - Do not reject for lack of novelty, significance or state-of-the-art results; judge only whether claims are supported and whether some audience would care. - If over-claiming is the problem, say which claims to scope down rather than rejecting the work outright. - Mark each requested change as critical to acceptance or as merely strengthening the paper. - Respectful, engaged tone; reviews become public on OpenReview. **Where the bar sits.** TMLR has no acceptance quota and publishes no rate. The action editor accepts when both criteria hold (claims supported, some audience interested), so a mean of 3 (Leaning Accept) or more is the rough line. Recommendation scores are mapped 1 Reject … 4 Accept by us; the labels are TMLR's. ## Reviewer form, field by field ### 1. Summary Of Contributions Type: free text. The paper's contributions in your own words, with its key strengths and weaknesses in brief. Not a substitute for the two criteria answers below. ### 2. Are the claims made in the submission supported by accurate, convincing and clear evidence? Type: single choice. The most important criterion. Judge only whether the stated claims are backed by accurate, convincing and clear evidence. Novelty and significance are explicitly NOT criteria; narrow claims that are fully supported are fine. - Yes - No ### 3. Claims And Evidence: Explanation Type: list of points. Explain the answer above: for each central claim, whether and how the evidence supports it. Unsupported or over-stated claims are the main reason to reject; say how they could be scoped down or supported. ### 4. Would at least some individuals in TMLR's audience be interested in knowing the findings of this paper? Type: single choice. A low bar: if unsure whether anyone would be interested, assume they would. Answer No only for findings no ML researcher or practitioner would care about. - Yes - No ### 5. Audience: Explanation Type: free text. Who in TMLR's audience would find the findings useful, and why (or why no one would). ### 6. Requested Changes Type: list of points. Proposed adjustments to the submission. Start each with "[Critical]" if it is critical to securing your recommendation for acceptance, or "[Strengthening]" if it would simply strengthen the work. ### 7. Broader Impact Concerns Type: free text. Ethical concerns that would require adding a Broader Impact Statement, or that the existing statement does not adequately address. "None" if there are none. ### 8. Recommendation Type: score, 1–4. Submitted on the separate Official Recommendation form after discussion. Must be consistent with the two criteria: Accept or Leaning Accept if and only if both answers are Yes. - **4**: Accept - **3**: Leaning Accept - **2**: Leaning Reject - **1**: Reject ## Area chair meta-review TMLR action editors decide from the two criteria, not from novelty. The decision has three outcomes: Accept as is, Accept with minor revision, or Reject (optionally inviting a major-revision resubmission). Recommendation options: Accept as is · Accept with minor revision · Reject ## LLM policy TMLR lets reviewers use LLMs and other tools to assist, e.g. to look up notation or polish grammar, but reviewers must form their own understanding of the paper and take responsibility for their reviews. Largely LLM-generated reviews are closely examined. Official policy: https://jmlr.org/tmlr/reviewer-guide.html --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=tmlr (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # COLT 2026 review form > Simulate COLT 2026-style learning-theory reviews: summary of results, significance and novelty relative to known bounds, correctness of the proofs, clarity, comments, an overall score and confidence. - Venue: Conference on Learning Theory - Form status: modeled, not official. COLT reviews in a private CMT form and publishes no reviewer guidelines. Every field and scale here is modeled on standard learning-theory reviewing (contribution, correctness, significance, clarity) and the CFP's reviewing philosophy. - Official source: https://learningtheory.org/colt2026/cfp.html - Overall score: Overall Score, 1–5 - Rough acceptance bar (our simulation): mean ≥ 3.5 - Web page: https://reviewpaper.ai/venues/colt ## How reviewers read the paper COLT asks reviewers to judge whether the work is close to being ready for publication and to help guide it there with constructive feedback. Papers are evaluated primarily on the main text, though all proofs must be included; there is an author response and discussion phase. Theory papers are judged on the significance and novelty of their results and techniques and on technical correctness, not on empirical evaluation. - Constructive: aim to help the paper reach publication quality. - Compare results precisely to the best known bounds and cite the prior work. - Say exactly which proofs were checked; point to specific steps for any error. - Do not ask for experiments unless the paper's claims are empirical. **Where the bar sits.** COLT 2026 accepted about 190 papers; submission counts are not published (historically ~30% acceptance). On this modeled 1–5 scale a mean of ≥ 3.5 is a rough proxy for acceptance. ## Reviewer form, field by field ### 1. Summary Of Results Type: free text. State the main theorems informally: setting, assumptions, the bounds or characterizations proved, and how they compare to the best known results. ### 2. Significance And Novelty Type: list of points. What is new relative to prior bounds and techniques (cite them), whether the techniques are of independent interest, and how the results advance learning theory. ### 3. Technical Correctness Type: list of points. Which proofs were checked and how carefully; any gaps, errors or unjustified steps (with locations); whether assumptions are reasonable and stated. Judgment rests primarily on the main text, but proofs must be complete in the appendix. ### 4. Clarity And Presentation Type: free text. Are the setting, results and proof ideas explained clearly in the main text? ### 5. Comments And Questions For The Authors Type: list of points. Constructive suggestions and questions for the author response, aimed at bringing the paper closer to publication. ### 6. Overall Score Type: score, 1–5. Overall recommendation. - **5**: Strong accept - **4**: Accept - **3**: Borderline - **2**: Reject - **1**: Strong reject ### 7. Confidence Type: score, 1–5. How certain you are, including how carefully the proofs were checked. - **5**: Expert: checked the proofs carefully and know the related literature well - **4**: Confident: checked the key proofs - **3**: Fairly confident: read the proofs at a high level - **2**: Somewhat confident: did not check the proofs - **1**: Educated guess ## Area chair meta-review COLT area chairs lead discussion and recommend; the form is not public. Recommendation options: Accept · Reject ## COLT submission requirements 1. At most 12 PMLR-format pages, plus unlimited references and appendices 2. All proofs included (main text should convey the key ideas) 3. Anonymized (double-blind) ## LLM policy COLT 2026 reviewers must not share submissions with any LLM; models may be used like books or Wikipedia to understand technical points, without leaking confidential information. Official policy: https://learningtheory.org/colt2026/cfp.html --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=colt (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # ACL ARR 2026 review form > Simulate ACL Rolling Review reviews for ACL, EMNLP and NAACL: Paper Summary, Strengths, numbered Weaknesses, Soundness and Excitement (1–5 in half points), Overall Assessment (Award / Conference / Findings / Resubmit), and the Responsible NLP checklist. - Venue: ACL Rolling Review (ACL, EMNLP, NAACL, EACL) - Form status: official (from the venue's public reviewer form) - Official source: https://aclrollingreview.org/reviewform - Overall score: Overall Assessment, 1–5 - Rough acceptance bar (our simulation): mean ≥ 3.5 - Web page: https://reviewpaper.ai/venues/acl-rolling-review ## How reviewers read the paper ARR reviews are about half a page. The form separates objective Soundness (must be justified in the text) from subjective Excitement, and the Overall Assessment is an outcome recommendation (Award / Conference / Findings / Resubmit), all on 1–5 with half points. Weaknesses are a numbered list specific enough for an AC to follow; suggested extra work goes under Comments. The guidelines list lazy-reviewing heuristics to avoid, e.g. "not novel" without a reference, "not SOTA", "simple method", "only tested on English", "I would have done X", or listing limitations the authors already acknowledge. - Number the weaknesses; each must be concrete enough for an AC to follow without reading the paper. - Avoid ARR lazy-reviewing heuristics: no "not novel" without citing the prior work, no "not SOTA" as a reason to reject, no "the method is too simple", no "only tested on English" unless the claims are general, no "I would have done X", no listing limitations the authors already acknowledge as weaknesses. - Put suggested additional experiments under Comments, not Weaknesses. - Justify the Soundness score in the text; Excitement is explicitly personal taste. **Where the bar sits.** ARR publishes no cutoff; venue senior ACs decide. Rule of thumb: average Overall Assessment ≥ 3.5 → main conference plausible, 3.0–3.5 → Findings, < 3.0 → likely rejected. Main conference acceptance is about 20–25%, plus ~13–15% Findings. ## Reviewer form, field by field ### 1. Paper Summary Type: free text. Describe what the paper is about so the AC and action editor can catch any misunderstanding. ### 2. Summary Of Strengths Type: list of points. The major reasons to publish this paper at a selective *ACL venue. ### 3. Summary Of Weaknesses Type: list of points. Numbered list of the concerns that would make you prioritize other high-quality papers. Each must be specific enough for an AC to follow without reading the paper. Do NOT include suggested extra work here (that goes in Comments), and do not list limitations the authors already acknowledge. ### 4. Comments Suggestions And Typos Type: list of points. Other improvements that are not part of the concerns above, including suggested additional experiments, presentation fixes and typos. ### 5. Confidence Type: score, 1–5. How confident you are in your evaluation. - **5**: Positive that my evaluation is correct. I read the paper very carefully and am familiar with related work. - **4**: Quite sure. I tried to check the important points carefully. It's unlikely, though conceivable, that I missed something that should affect my ratings. - **3**: Pretty sure, but there's a chance I missed something. Although I have a good feel for this area in general, I did not carefully check the paper's details, e.g., the math or experimental design. - **2**: Willing to defend my evaluation, but it is fairly likely that I missed some details, didn't understand some central points, or can't be sure about the novelty of the work. - **1**: Not my area, or paper is very hard to understand. My evaluation is just an educated guess. ### 6. Soundness Type: score, 1–5 (step 0.5). Objective: does the study support its claims? Must be justified in the review text. - **5**: Excellent: This study is one of the most thorough I have seen, given its type. - **4**: Strong: This study provides sufficient support for all of its claims. Some extra experiments could be nice, but not essential. - **3**: Acceptable: This study provides sufficient support for its main claims. Some minor points may need extra support or details. - **2**: Poor: Some of the main claims are not sufficiently supported. There are major technical/methodological problems. - **1**: Major Issues: This study is not yet sufficiently thorough to warrant publication or is not relevant to ACL. ### 7. Excitement Type: score, 1–5 (step 0.5). Explicitly subjective: how exciting you personally find the work. - **5**: Highly Exciting: I would recommend this paper to others and/or attend its presentation in a conference. - **4**: Exciting: I would mention this paper to others and/or make an effort to attend its presentation in a conference. - **3**: Interesting: I might mention some points of this paper to others and/or attend its presentation in a conference if there's time. - **2**: Potentially Interesting: this paper does not resonate with me, but it might with others in the *ACL community. - **1**: Not Exciting: this paper does not resonate with me, and I don't think it would with others in the *ACL community. ### 8. Overall Assessment Type: score, 1–5 (step 0.5). An outcome recommendation (Award / Conference / Findings / Resubmit), not a quality score. - **5**: Consider for Award: I think this paper could be considered for an outstanding paper award at an *ACL conference (up to top 2.5% papers). - **4.5**: Borderline Award - **4**: Conference: I think this paper could be accepted to an *ACL conference. - **3.5**: Borderline Conference - **3**: Findings: I think this paper could be accepted to the Findings of the ACL. - **2.5**: Borderline Findings - **2**: Resubmit next cycle: I think this paper needs substantial revisions that can be completed by the next ARR cycle. - **1.5**: Resubmit after next cycle: I think this paper needs substantial revisions that cannot be completed by the next ARR cycle. - **1**: Do not resubmit: This paper has to be fully redone, or it is not relevant to the *ACL community (e.g., it is in no way related to computational processing of language). ### 9. Best Paper Justification (optional) Type: free text. Only if Overall Assessment is 4.5 or 5; otherwise write "N/A". ### 10. Limitations And Societal Impact Type: free text. Have the authors adequately discussed the limitations and potential positive and negative societal impacts of their work? ### 11. Ethical Concerns Type: free text. Concerns with respect to the ACL Code of Ethics and the Responsible NLP checklist. "None" if there are none. ### 12. Needs Ethics Review Type: single choice. Should the paper go to the ethics committee? Answer No or Yes (with a reason). - No - Yes ### 13. Reproducibility Type: score, 1–5. Could others reproduce the results? - **5**: They could easily reproduce the results. - **4**: They could mostly reproduce the results, but there may be some variation because of sample variance or minor variations in their interpretation of the protocol or method. - **3**: They could reproduce the results with some difficulty. The settings of parameters are underspecified or subjectively determined, and/or the training/evaluation data are not widely available. - **2**: They would be hard pressed to reproduce the results. The contribution depends on data that are simply not available outside the author's institution or consortium and/or not enough details are provided. - **1**: They would not be able to reproduce the results here no matter how hard they tried. ### 14. Datasets Type: score, 1–5. Value of any newly released datasets (1 if none). - **5**: Enabling: The newly released datasets should affect other people's choice of research or development projects to undertake. - **4**: Useful: I would recommend the new datasets to other researchers or developers for their ongoing work. - **3**: Potentially useful: Someone might find the new datasets useful for their work. - **2**: Documentary: The new datasets will be useful to study or replicate the reported research, although for other purposes they may have limited interest or limited usability. (Still a positive rating) - **1**: No usable datasets submitted. ### 15. Software Type: score, 1–5. Value of any newly released software (1 if none). - **5**: Enabling: The newly released software should affect other people's choice of research or development projects to undertake. - **4**: Useful: I would recommend the new software to other researchers or developers for their ongoing work. - **3**: Potentially useful: Someone might find the new software useful for their work. - **2**: Documentary: The new software will be useful to study or replicate the reported research, although for other purposes they may have limited interest or limited usability. (Still a positive rating) - **1**: No usable software released. ## Area chair meta-review ARR meta-reviews use the reviewers' Overall Assessment scale. Weigh arguments, not averages. Recommendation options: Conference · Findings · Resubmit next cycle · Resubmit after next cycle · Do not resubmit ## Responsible NLP Research Checklist 1. A1. Limitations section describing the work's limitations (required; missing it is grounds for desk rejection) 2. A2. Potential risks of the work discussed 3. B1. Creators of artifacts used are cited 4. B2. License and terms of use of artifacts discussed 5. B3. Use of existing artifacts consistent with their intended use 6. B4. Data checked for personally identifying or offensive content and anonymized 7. B5. Documentation of artifacts (domains, languages, linguistic phenomena, demographics) 8. B6. Relevant statistics (number of examples, train/test/dev splits) reported 9. C1. Number of parameters, total computational budget and computing infrastructure reported 10. C2. Experimental setup, hyperparameter search and best-found values discussed 11. C3. Descriptive statistics (e.g. error bars) reported, and whether results are single runs or aggregates 12. C4. Implementation, model and parameter settings of existing packages reported 13. D1. Full text of instructions given to human participants/annotators 14. D2. Recruitment and payment of participants, and whether payment was adequate 15. D3. Consent obtained from participants whose data is used 16. D4. Data collection protocol approved by an ethics review board 17. D5. Basic demographic and geographic characteristics of annotators reported 18. E1. Use of AI assistants (in research or writing) disclosed ## LLM policy ARR forbids using generative AI to create the first draft of a review or meta-review. Light language editing is allowed, but manuscripts and reviews must not be uploaded to non-privacy-preserving tools. Official policy: https://aclrollingreview.org/reviewerguidelines --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=acl-rolling-review (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # EMNLP 2026 review form > Simulate EMNLP 2026 reviews in the ACL Rolling Review form: Paper Summary, Strengths, numbered Weaknesses, Soundness and Excitement (1–5 in half points), and the Overall Assessment that routes papers to EMNLP main, Findings or resubmission. - Venue: Conference on Empirical Methods in Natural Language Processing - Form status: official (from the venue's public reviewer form) - Official source: https://aclrollingreview.org/reviewform - Overall score: Overall Assessment, 1–5 - Rough acceptance bar (our simulation): mean ≥ 3.5 - Web page: https://reviewpaper.ai/venues/emnlp ## How reviewers read the paper EMNLP 2026 does not run its own reviewing: papers are reviewed in ACL Rolling Review and then committed to EMNLP, whose program chairs decide from the ARR reviews and meta-review. EMNLP 2026 takes papers from the ARR May 2026 cycle (or earlier cycles); the target conference declared at ARR submission is binding, and commitment closed on 2 August 2026. ARR reviews are about half a page. The form separates objective Soundness (must be justified in the text) from subjective Excitement, and the Overall Assessment is an outcome recommendation (Award / Conference / Findings / Resubmit), all on 1–5 with half points. Weaknesses are a numbered list specific enough for an AC to follow; suggested extra work goes under Comments. The guidelines list lazy-reviewing heuristics to avoid, e.g. "not novel" without a reference, "not SOTA", "simple method", "only tested on English", "I would have done X", or listing limitations the authors already acknowledge. - Number the weaknesses; each must be concrete enough for an AC to follow without reading the paper. - Avoid ARR lazy-reviewing heuristics: no "not novel" without citing the prior work, no "not SOTA" as a reason to reject, no "the method is too simple", no "only tested on English" unless the claims are general, no "I would have done X", no listing limitations the authors already acknowledge as weaknesses. - Put suggested additional experiments under Comments, not Weaknesses. - Justify the Soundness score in the text; Excitement is explicitly personal taste. **Where the bar sits.** EMNLP has published a Findings of EMNLP volume every year since 2020. In the Overall Assessment, 4 ("Conference") argues for the main conference and 3 ("Findings") for the Findings of the ACL volume, which the program chairs may offer to committed papers that are sound but less exciting. ARR publishes no cutoff; venue senior ACs decide. Rule of thumb: average Overall Assessment ≥ 3.5 → main conference plausible, 3.0–3.5 → Findings, < 3.0 → likely rejected. Main conference acceptance is about 20–25%, plus ~13–15% Findings. ## Reviewer form, field by field ### 1. Paper Summary Type: free text. Describe what the paper is about so the AC and action editor can catch any misunderstanding. ### 2. Summary Of Strengths Type: list of points. The major reasons to publish this paper at a selective *ACL venue. ### 3. Summary Of Weaknesses Type: list of points. Numbered list of the concerns that would make you prioritize other high-quality papers. Each must be specific enough for an AC to follow without reading the paper. Do NOT include suggested extra work here (that goes in Comments), and do not list limitations the authors already acknowledge. ### 4. Comments Suggestions And Typos Type: list of points. Other improvements that are not part of the concerns above, including suggested additional experiments, presentation fixes and typos. ### 5. Confidence Type: score, 1–5. How confident you are in your evaluation. - **5**: Positive that my evaluation is correct. I read the paper very carefully and am familiar with related work. - **4**: Quite sure. I tried to check the important points carefully. It's unlikely, though conceivable, that I missed something that should affect my ratings. - **3**: Pretty sure, but there's a chance I missed something. Although I have a good feel for this area in general, I did not carefully check the paper's details, e.g., the math or experimental design. - **2**: Willing to defend my evaluation, but it is fairly likely that I missed some details, didn't understand some central points, or can't be sure about the novelty of the work. - **1**: Not my area, or paper is very hard to understand. My evaluation is just an educated guess. ### 6. Soundness Type: score, 1–5 (step 0.5). Objective: does the study support its claims? Must be justified in the review text. - **5**: Excellent: This study is one of the most thorough I have seen, given its type. - **4**: Strong: This study provides sufficient support for all of its claims. Some extra experiments could be nice, but not essential. - **3**: Acceptable: This study provides sufficient support for its main claims. Some minor points may need extra support or details. - **2**: Poor: Some of the main claims are not sufficiently supported. There are major technical/methodological problems. - **1**: Major Issues: This study is not yet sufficiently thorough to warrant publication or is not relevant to ACL. ### 7. Excitement Type: score, 1–5 (step 0.5). Explicitly subjective: how exciting you personally find the work. - **5**: Highly Exciting: I would recommend this paper to others and/or attend its presentation in a conference. - **4**: Exciting: I would mention this paper to others and/or make an effort to attend its presentation in a conference. - **3**: Interesting: I might mention some points of this paper to others and/or attend its presentation in a conference if there's time. - **2**: Potentially Interesting: this paper does not resonate with me, but it might with others in the *ACL community. - **1**: Not Exciting: this paper does not resonate with me, and I don't think it would with others in the *ACL community. ### 8. Overall Assessment Type: score, 1–5 (step 0.5). An outcome recommendation (Award / Conference / Findings / Resubmit), not a quality score. - **5**: Consider for Award: I think this paper could be considered for an outstanding paper award at an *ACL conference (up to top 2.5% papers). - **4.5**: Borderline Award - **4**: Conference: I think this paper could be accepted to an *ACL conference. - **3.5**: Borderline Conference - **3**: Findings: I think this paper could be accepted to the Findings of the ACL. - **2.5**: Borderline Findings - **2**: Resubmit next cycle: I think this paper needs substantial revisions that can be completed by the next ARR cycle. - **1.5**: Resubmit after next cycle: I think this paper needs substantial revisions that cannot be completed by the next ARR cycle. - **1**: Do not resubmit: This paper has to be fully redone, or it is not relevant to the *ACL community (e.g., it is in no way related to computational processing of language). ### 9. Best Paper Justification (optional) Type: free text. Only if Overall Assessment is 4.5 or 5; otherwise write "N/A". ### 10. Limitations And Societal Impact Type: free text. Have the authors adequately discussed the limitations and potential positive and negative societal impacts of their work? ### 11. Ethical Concerns Type: free text. Concerns with respect to the ACL Code of Ethics and the Responsible NLP checklist. "None" if there are none. ### 12. Needs Ethics Review Type: single choice. Should the paper go to the ethics committee? Answer No or Yes (with a reason). - No - Yes ### 13. Reproducibility Type: score, 1–5. Could others reproduce the results? - **5**: They could easily reproduce the results. - **4**: They could mostly reproduce the results, but there may be some variation because of sample variance or minor variations in their interpretation of the protocol or method. - **3**: They could reproduce the results with some difficulty. The settings of parameters are underspecified or subjectively determined, and/or the training/evaluation data are not widely available. - **2**: They would be hard pressed to reproduce the results. The contribution depends on data that are simply not available outside the author's institution or consortium and/or not enough details are provided. - **1**: They would not be able to reproduce the results here no matter how hard they tried. ### 14. Datasets Type: score, 1–5. Value of any newly released datasets (1 if none). - **5**: Enabling: The newly released datasets should affect other people's choice of research or development projects to undertake. - **4**: Useful: I would recommend the new datasets to other researchers or developers for their ongoing work. - **3**: Potentially useful: Someone might find the new datasets useful for their work. - **2**: Documentary: The new datasets will be useful to study or replicate the reported research, although for other purposes they may have limited interest or limited usability. (Still a positive rating) - **1**: No usable datasets submitted. ### 15. Software Type: score, 1–5. Value of any newly released software (1 if none). - **5**: Enabling: The newly released software should affect other people's choice of research or development projects to undertake. - **4**: Useful: I would recommend the new software to other researchers or developers for their ongoing work. - **3**: Potentially useful: Someone might find the new software useful for their work. - **2**: Documentary: The new software will be useful to study or replicate the reported research, although for other purposes they may have limited interest or limited usability. (Still a positive rating) - **1**: No usable software released. ## Area chair meta-review ARR meta-reviews use the reviewers' Overall Assessment scale. Weigh arguments, not averages. Recommendation options: Conference · Findings · Resubmit next cycle · Resubmit after next cycle · Do not resubmit ## Responsible NLP Research Checklist 1. A1. Limitations section describing the work's limitations (required; missing it is grounds for desk rejection) 2. A2. Potential risks of the work discussed 3. B1. Creators of artifacts used are cited 4. B2. License and terms of use of artifacts discussed 5. B3. Use of existing artifacts consistent with their intended use 6. B4. Data checked for personally identifying or offensive content and anonymized 7. B5. Documentation of artifacts (domains, languages, linguistic phenomena, demographics) 8. B6. Relevant statistics (number of examples, train/test/dev splits) reported 9. C1. Number of parameters, total computational budget and computing infrastructure reported 10. C2. Experimental setup, hyperparameter search and best-found values discussed 11. C3. Descriptive statistics (e.g. error bars) reported, and whether results are single runs or aggregates 12. C4. Implementation, model and parameter settings of existing packages reported 13. D1. Full text of instructions given to human participants/annotators 14. D2. Recruitment and payment of participants, and whether payment was adequate 15. D3. Consent obtained from participants whose data is used 16. D4. Data collection protocol approved by an ethics review board 17. D5. Basic demographic and geographic characteristics of annotators reported 18. E1. Use of AI assistants (in research or writing) disclosed ## LLM policy ARR forbids using generative AI to create the first draft of a review or meta-review. Light language editing is allowed, but manuscripts and reviews must not be uploaded to non-privacy-preserving tools. Official policy: https://aclrollingreview.org/reviewerguidelines --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=emnlp (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # NAACL 2027 review form > Simulate NAACL 2027 reviews in the ACL Rolling Review form: Paper Summary, Strengths, numbered Weaknesses, Soundness and Excitement (1–5 in half points), and the Overall Assessment that routes papers to NAACL main, Findings or resubmission. - Venue: Annual Conference of the Nations of the Americas Chapter of the ACL - Form status: official (from the venue's public reviewer form) - Official source: https://aclrollingreview.org/reviewform - Overall score: Overall Assessment, 1–5 - Rough acceptance bar (our simulation): mean ≥ 3.5 - Web page: https://reviewpaper.ai/venues/naacl ## How reviewers read the paper NAACL 2027 does not run its own reviewing: papers are reviewed in ACL Rolling Review and then committed to NAACL, whose program chairs decide from the ARR reviews and meta-review. NAACL 2027 (San Francisco, June 2027) takes papers from the ARR October 2026 cycle, shared with COLING 2027; at commitment (deadline 23 December 2026) authors pick one primary venue and may opt in to be considered by the other. ARR reviews are about half a page. The form separates objective Soundness (must be justified in the text) from subjective Excitement, and the Overall Assessment is an outcome recommendation (Award / Conference / Findings / Resubmit), all on 1–5 with half points. Weaknesses are a numbered list specific enough for an AC to follow; suggested extra work goes under Comments. The guidelines list lazy-reviewing heuristics to avoid, e.g. "not novel" without a reference, "not SOTA", "simple method", "only tested on English", "I would have done X", or listing limitations the authors already acknowledge. - Number the weaknesses; each must be concrete enough for an AC to follow without reading the paper. - Avoid ARR lazy-reviewing heuristics: no "not novel" without citing the prior work, no "not SOTA" as a reason to reject, no "the method is too simple", no "only tested on English" unless the claims are general, no "I would have done X", no listing limitations the authors already acknowledge as weaknesses. - Put suggested additional experiments under Comments, not Weaknesses. - Justify the Soundness score in the text; Excitement is explicitly personal taste. **Where the bar sits.** Recent NAACL editions (2022, 2024, 2025) published a Findings volume. In the Overall Assessment, 4 ("Conference") argues for the main conference and 3 ("Findings") for the Findings of the ACL volume, which the program chairs may offer to committed papers that are sound but less exciting. ARR publishes no cutoff; venue senior ACs decide. Rule of thumb: average Overall Assessment ≥ 3.5 → main conference plausible, 3.0–3.5 → Findings, < 3.0 → likely rejected. Main conference acceptance is about 20–25%, plus ~13–15% Findings. ## Reviewer form, field by field ### 1. Paper Summary Type: free text. Describe what the paper is about so the AC and action editor can catch any misunderstanding. ### 2. Summary Of Strengths Type: list of points. The major reasons to publish this paper at a selective *ACL venue. ### 3. Summary Of Weaknesses Type: list of points. Numbered list of the concerns that would make you prioritize other high-quality papers. Each must be specific enough for an AC to follow without reading the paper. Do NOT include suggested extra work here (that goes in Comments), and do not list limitations the authors already acknowledge. ### 4. Comments Suggestions And Typos Type: list of points. Other improvements that are not part of the concerns above, including suggested additional experiments, presentation fixes and typos. ### 5. Confidence Type: score, 1–5. How confident you are in your evaluation. - **5**: Positive that my evaluation is correct. I read the paper very carefully and am familiar with related work. - **4**: Quite sure. I tried to check the important points carefully. It's unlikely, though conceivable, that I missed something that should affect my ratings. - **3**: Pretty sure, but there's a chance I missed something. Although I have a good feel for this area in general, I did not carefully check the paper's details, e.g., the math or experimental design. - **2**: Willing to defend my evaluation, but it is fairly likely that I missed some details, didn't understand some central points, or can't be sure about the novelty of the work. - **1**: Not my area, or paper is very hard to understand. My evaluation is just an educated guess. ### 6. Soundness Type: score, 1–5 (step 0.5). Objective: does the study support its claims? Must be justified in the review text. - **5**: Excellent: This study is one of the most thorough I have seen, given its type. - **4**: Strong: This study provides sufficient support for all of its claims. Some extra experiments could be nice, but not essential. - **3**: Acceptable: This study provides sufficient support for its main claims. Some minor points may need extra support or details. - **2**: Poor: Some of the main claims are not sufficiently supported. There are major technical/methodological problems. - **1**: Major Issues: This study is not yet sufficiently thorough to warrant publication or is not relevant to ACL. ### 7. Excitement Type: score, 1–5 (step 0.5). Explicitly subjective: how exciting you personally find the work. - **5**: Highly Exciting: I would recommend this paper to others and/or attend its presentation in a conference. - **4**: Exciting: I would mention this paper to others and/or make an effort to attend its presentation in a conference. - **3**: Interesting: I might mention some points of this paper to others and/or attend its presentation in a conference if there's time. - **2**: Potentially Interesting: this paper does not resonate with me, but it might with others in the *ACL community. - **1**: Not Exciting: this paper does not resonate with me, and I don't think it would with others in the *ACL community. ### 8. Overall Assessment Type: score, 1–5 (step 0.5). An outcome recommendation (Award / Conference / Findings / Resubmit), not a quality score. - **5**: Consider for Award: I think this paper could be considered for an outstanding paper award at an *ACL conference (up to top 2.5% papers). - **4.5**: Borderline Award - **4**: Conference: I think this paper could be accepted to an *ACL conference. - **3.5**: Borderline Conference - **3**: Findings: I think this paper could be accepted to the Findings of the ACL. - **2.5**: Borderline Findings - **2**: Resubmit next cycle: I think this paper needs substantial revisions that can be completed by the next ARR cycle. - **1.5**: Resubmit after next cycle: I think this paper needs substantial revisions that cannot be completed by the next ARR cycle. - **1**: Do not resubmit: This paper has to be fully redone, or it is not relevant to the *ACL community (e.g., it is in no way related to computational processing of language). ### 9. Best Paper Justification (optional) Type: free text. Only if Overall Assessment is 4.5 or 5; otherwise write "N/A". ### 10. Limitations And Societal Impact Type: free text. Have the authors adequately discussed the limitations and potential positive and negative societal impacts of their work? ### 11. Ethical Concerns Type: free text. Concerns with respect to the ACL Code of Ethics and the Responsible NLP checklist. "None" if there are none. ### 12. Needs Ethics Review Type: single choice. Should the paper go to the ethics committee? Answer No or Yes (with a reason). - No - Yes ### 13. Reproducibility Type: score, 1–5. Could others reproduce the results? - **5**: They could easily reproduce the results. - **4**: They could mostly reproduce the results, but there may be some variation because of sample variance or minor variations in their interpretation of the protocol or method. - **3**: They could reproduce the results with some difficulty. The settings of parameters are underspecified or subjectively determined, and/or the training/evaluation data are not widely available. - **2**: They would be hard pressed to reproduce the results. The contribution depends on data that are simply not available outside the author's institution or consortium and/or not enough details are provided. - **1**: They would not be able to reproduce the results here no matter how hard they tried. ### 14. Datasets Type: score, 1–5. Value of any newly released datasets (1 if none). - **5**: Enabling: The newly released datasets should affect other people's choice of research or development projects to undertake. - **4**: Useful: I would recommend the new datasets to other researchers or developers for their ongoing work. - **3**: Potentially useful: Someone might find the new datasets useful for their work. - **2**: Documentary: The new datasets will be useful to study or replicate the reported research, although for other purposes they may have limited interest or limited usability. (Still a positive rating) - **1**: No usable datasets submitted. ### 15. Software Type: score, 1–5. Value of any newly released software (1 if none). - **5**: Enabling: The newly released software should affect other people's choice of research or development projects to undertake. - **4**: Useful: I would recommend the new software to other researchers or developers for their ongoing work. - **3**: Potentially useful: Someone might find the new software useful for their work. - **2**: Documentary: The new software will be useful to study or replicate the reported research, although for other purposes they may have limited interest or limited usability. (Still a positive rating) - **1**: No usable software released. ## Area chair meta-review ARR meta-reviews use the reviewers' Overall Assessment scale. Weigh arguments, not averages. Recommendation options: Conference · Findings · Resubmit next cycle · Resubmit after next cycle · Do not resubmit ## Responsible NLP Research Checklist 1. A1. Limitations section describing the work's limitations (required; missing it is grounds for desk rejection) 2. A2. Potential risks of the work discussed 3. B1. Creators of artifacts used are cited 4. B2. License and terms of use of artifacts discussed 5. B3. Use of existing artifacts consistent with their intended use 6. B4. Data checked for personally identifying or offensive content and anonymized 7. B5. Documentation of artifacts (domains, languages, linguistic phenomena, demographics) 8. B6. Relevant statistics (number of examples, train/test/dev splits) reported 9. C1. Number of parameters, total computational budget and computing infrastructure reported 10. C2. Experimental setup, hyperparameter search and best-found values discussed 11. C3. Descriptive statistics (e.g. error bars) reported, and whether results are single runs or aggregates 12. C4. Implementation, model and parameter settings of existing packages reported 13. D1. Full text of instructions given to human participants/annotators 14. D2. Recruitment and payment of participants, and whether payment was adequate 15. D3. Consent obtained from participants whose data is used 16. D4. Data collection protocol approved by an ethics review board 17. D5. Basic demographic and geographic characteristics of annotators reported 18. E1. Use of AI assistants (in research or writing) disclosed ## LLM policy ARR forbids using generative AI to create the first draft of a review or meta-review. Light language editing is allowed, but manuscripts and reviews must not be uploaded to non-privacy-preserving tools. Official policy: https://aclrollingreview.org/reviewerguidelines --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=naacl (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # COLING 2027 review form > Simulate COLING 2027 reviews in the ACL Rolling Review form: Paper Summary, Strengths, numbered Weaknesses, Soundness and Excitement (1–5 in half points), and the Overall Assessment that routes papers to COLING main, Findings or resubmission. - Venue: International Conference on Computational Linguistics - Form status: official (from the venue's public reviewer form) - Official source: https://aclrollingreview.org/reviewform - Overall score: Overall Assessment, 1–5 - Rough acceptance bar (our simulation): mean ≥ 3.5 - Web page: https://reviewpaper.ai/venues/coling ## How reviewers read the paper COLING 2027 does not run its own reviewing: papers are reviewed in ACL Rolling Review and then committed to COLING, whose program chairs decide from the ARR reviews and meta-review. COLING 2027 takes papers from the ARR October 2026 cycle, shared with NAACL 2027; at commitment authors pick one primary venue and may opt in to be considered by the other. ARR reviews are about half a page. The form separates objective Soundness (must be justified in the text) from subjective Excitement, and the Overall Assessment is an outcome recommendation (Award / Conference / Findings / Resubmit), all on 1–5 with half points. Weaknesses are a numbered list specific enough for an AC to follow; suggested extra work goes under Comments. The guidelines list lazy-reviewing heuristics to avoid, e.g. "not novel" without a reference, "not SOTA", "simple method", "only tested on English", "I would have done X", or listing limitations the authors already acknowledge. - Number the weaknesses; each must be concrete enough for an AC to follow without reading the paper. - Avoid ARR lazy-reviewing heuristics: no "not novel" without citing the prior work, no "not SOTA" as a reason to reject, no "the method is too simple", no "only tested on English" unless the claims are general, no "I would have done X", no listing limitations the authors already acknowledge as weaknesses. - Put suggested additional experiments under Comments, not Weaknesses. - Justify the Soundness score in the text; Excitement is explicitly personal taste. **Where the bar sits.** COLING has not historically published a Findings volume; check the COLING 2027 call. In the Overall Assessment, 4 ("Conference") argues for the main conference and 3 ("Findings") for the Findings of the ACL volume, which the program chairs may offer to committed papers that are sound but less exciting. ARR publishes no cutoff; venue senior ACs decide. Rule of thumb: average Overall Assessment ≥ 3.5 → main conference plausible, 3.0–3.5 → Findings, < 3.0 → likely rejected. Main conference acceptance is about 20–25%, plus ~13–15% Findings. ## Reviewer form, field by field ### 1. Paper Summary Type: free text. Describe what the paper is about so the AC and action editor can catch any misunderstanding. ### 2. Summary Of Strengths Type: list of points. The major reasons to publish this paper at a selective *ACL venue. ### 3. Summary Of Weaknesses Type: list of points. Numbered list of the concerns that would make you prioritize other high-quality papers. Each must be specific enough for an AC to follow without reading the paper. Do NOT include suggested extra work here (that goes in Comments), and do not list limitations the authors already acknowledge. ### 4. Comments Suggestions And Typos Type: list of points. Other improvements that are not part of the concerns above, including suggested additional experiments, presentation fixes and typos. ### 5. Confidence Type: score, 1–5. How confident you are in your evaluation. - **5**: Positive that my evaluation is correct. I read the paper very carefully and am familiar with related work. - **4**: Quite sure. I tried to check the important points carefully. It's unlikely, though conceivable, that I missed something that should affect my ratings. - **3**: Pretty sure, but there's a chance I missed something. Although I have a good feel for this area in general, I did not carefully check the paper's details, e.g., the math or experimental design. - **2**: Willing to defend my evaluation, but it is fairly likely that I missed some details, didn't understand some central points, or can't be sure about the novelty of the work. - **1**: Not my area, or paper is very hard to understand. My evaluation is just an educated guess. ### 6. Soundness Type: score, 1–5 (step 0.5). Objective: does the study support its claims? Must be justified in the review text. - **5**: Excellent: This study is one of the most thorough I have seen, given its type. - **4**: Strong: This study provides sufficient support for all of its claims. Some extra experiments could be nice, but not essential. - **3**: Acceptable: This study provides sufficient support for its main claims. Some minor points may need extra support or details. - **2**: Poor: Some of the main claims are not sufficiently supported. There are major technical/methodological problems. - **1**: Major Issues: This study is not yet sufficiently thorough to warrant publication or is not relevant to ACL. ### 7. Excitement Type: score, 1–5 (step 0.5). Explicitly subjective: how exciting you personally find the work. - **5**: Highly Exciting: I would recommend this paper to others and/or attend its presentation in a conference. - **4**: Exciting: I would mention this paper to others and/or make an effort to attend its presentation in a conference. - **3**: Interesting: I might mention some points of this paper to others and/or attend its presentation in a conference if there's time. - **2**: Potentially Interesting: this paper does not resonate with me, but it might with others in the *ACL community. - **1**: Not Exciting: this paper does not resonate with me, and I don't think it would with others in the *ACL community. ### 8. Overall Assessment Type: score, 1–5 (step 0.5). An outcome recommendation (Award / Conference / Findings / Resubmit), not a quality score. - **5**: Consider for Award: I think this paper could be considered for an outstanding paper award at an *ACL conference (up to top 2.5% papers). - **4.5**: Borderline Award - **4**: Conference: I think this paper could be accepted to an *ACL conference. - **3.5**: Borderline Conference - **3**: Findings: I think this paper could be accepted to the Findings of the ACL. - **2.5**: Borderline Findings - **2**: Resubmit next cycle: I think this paper needs substantial revisions that can be completed by the next ARR cycle. - **1.5**: Resubmit after next cycle: I think this paper needs substantial revisions that cannot be completed by the next ARR cycle. - **1**: Do not resubmit: This paper has to be fully redone, or it is not relevant to the *ACL community (e.g., it is in no way related to computational processing of language). ### 9. Best Paper Justification (optional) Type: free text. Only if Overall Assessment is 4.5 or 5; otherwise write "N/A". ### 10. Limitations And Societal Impact Type: free text. Have the authors adequately discussed the limitations and potential positive and negative societal impacts of their work? ### 11. Ethical Concerns Type: free text. Concerns with respect to the ACL Code of Ethics and the Responsible NLP checklist. "None" if there are none. ### 12. Needs Ethics Review Type: single choice. Should the paper go to the ethics committee? Answer No or Yes (with a reason). - No - Yes ### 13. Reproducibility Type: score, 1–5. Could others reproduce the results? - **5**: They could easily reproduce the results. - **4**: They could mostly reproduce the results, but there may be some variation because of sample variance or minor variations in their interpretation of the protocol or method. - **3**: They could reproduce the results with some difficulty. The settings of parameters are underspecified or subjectively determined, and/or the training/evaluation data are not widely available. - **2**: They would be hard pressed to reproduce the results. The contribution depends on data that are simply not available outside the author's institution or consortium and/or not enough details are provided. - **1**: They would not be able to reproduce the results here no matter how hard they tried. ### 14. Datasets Type: score, 1–5. Value of any newly released datasets (1 if none). - **5**: Enabling: The newly released datasets should affect other people's choice of research or development projects to undertake. - **4**: Useful: I would recommend the new datasets to other researchers or developers for their ongoing work. - **3**: Potentially useful: Someone might find the new datasets useful for their work. - **2**: Documentary: The new datasets will be useful to study or replicate the reported research, although for other purposes they may have limited interest or limited usability. (Still a positive rating) - **1**: No usable datasets submitted. ### 15. Software Type: score, 1–5. Value of any newly released software (1 if none). - **5**: Enabling: The newly released software should affect other people's choice of research or development projects to undertake. - **4**: Useful: I would recommend the new software to other researchers or developers for their ongoing work. - **3**: Potentially useful: Someone might find the new software useful for their work. - **2**: Documentary: The new software will be useful to study or replicate the reported research, although for other purposes they may have limited interest or limited usability. (Still a positive rating) - **1**: No usable software released. ## Area chair meta-review ARR meta-reviews use the reviewers' Overall Assessment scale. Weigh arguments, not averages. Recommendation options: Conference · Findings · Resubmit next cycle · Resubmit after next cycle · Do not resubmit ## Responsible NLP Research Checklist 1. A1. Limitations section describing the work's limitations (required; missing it is grounds for desk rejection) 2. A2. Potential risks of the work discussed 3. B1. Creators of artifacts used are cited 4. B2. License and terms of use of artifacts discussed 5. B3. Use of existing artifacts consistent with their intended use 6. B4. Data checked for personally identifying or offensive content and anonymized 7. B5. Documentation of artifacts (domains, languages, linguistic phenomena, demographics) 8. B6. Relevant statistics (number of examples, train/test/dev splits) reported 9. C1. Number of parameters, total computational budget and computing infrastructure reported 10. C2. Experimental setup, hyperparameter search and best-found values discussed 11. C3. Descriptive statistics (e.g. error bars) reported, and whether results are single runs or aggregates 12. C4. Implementation, model and parameter settings of existing packages reported 13. D1. Full text of instructions given to human participants/annotators 14. D2. Recruitment and payment of participants, and whether payment was adequate 15. D3. Consent obtained from participants whose data is used 16. D4. Data collection protocol approved by an ethics review board 17. D5. Basic demographic and geographic characteristics of annotators reported 18. E1. Use of AI assistants (in research or writing) disclosed ## LLM policy ARR forbids using generative AI to create the first draft of a review or meta-review. Light language editing is allowed, but manuscripts and reviews must not be uploaded to non-privacy-preserving tools. Official policy: https://aclrollingreview.org/reviewerguidelines --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=coling (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # EACL 2027 review form > Simulate EACL 2027 reviews in the ACL Rolling Review form: Paper Summary, Strengths, numbered Weaknesses, Soundness and Excitement (1–5 in half points), and the Overall Assessment that routes papers to EACL main, Findings or resubmission. - Venue: Conference of the European Chapter of the ACL - Form status: official (from the venue's public reviewer form) - Official source: https://aclrollingreview.org/reviewform - Overall score: Overall Assessment, 1–5 - Rough acceptance bar (our simulation): mean ≥ 3.5 - Web page: https://reviewpaper.ai/venues/eacl ## How reviewers read the paper EACL 2027 does not run its own reviewing: papers are reviewed in ACL Rolling Review and then committed to EACL, whose program chairs decide from the ARR reviews and meta-review. EACL 2027 (March 2027, special theme "The Human in Language") takes papers from the ARR August 2026 cycle or earlier; commitment closed on 11 October 2026. ARR reviews are about half a page. The form separates objective Soundness (must be justified in the text) from subjective Excitement, and the Overall Assessment is an outcome recommendation (Award / Conference / Findings / Resubmit), all on 1–5 with half points. Weaknesses are a numbered list specific enough for an AC to follow; suggested extra work goes under Comments. The guidelines list lazy-reviewing heuristics to avoid, e.g. "not novel" without a reference, "not SOTA", "simple method", "only tested on English", "I would have done X", or listing limitations the authors already acknowledge. - Number the weaknesses; each must be concrete enough for an AC to follow without reading the paper. - Avoid ARR lazy-reviewing heuristics: no "not novel" without citing the prior work, no "not SOTA" as a reason to reject, no "the method is too simple", no "only tested on English" unless the claims are general, no "I would have done X", no listing limitations the authors already acknowledge as weaknesses. - Put suggested additional experiments under Comments, not Weaknesses. - Justify the Soundness score in the text; Excitement is explicitly personal taste. **Where the bar sits.** EACL has published a Findings of EACL volume since 2023. In the Overall Assessment, 4 ("Conference") argues for the main conference and 3 ("Findings") for the Findings of the ACL volume, which the program chairs may offer to committed papers that are sound but less exciting. ARR publishes no cutoff; venue senior ACs decide. Rule of thumb: average Overall Assessment ≥ 3.5 → main conference plausible, 3.0–3.5 → Findings, < 3.0 → likely rejected. Main conference acceptance is about 20–25%, plus ~13–15% Findings. ## Reviewer form, field by field ### 1. Paper Summary Type: free text. Describe what the paper is about so the AC and action editor can catch any misunderstanding. ### 2. Summary Of Strengths Type: list of points. The major reasons to publish this paper at a selective *ACL venue. ### 3. Summary Of Weaknesses Type: list of points. Numbered list of the concerns that would make you prioritize other high-quality papers. Each must be specific enough for an AC to follow without reading the paper. Do NOT include suggested extra work here (that goes in Comments), and do not list limitations the authors already acknowledge. ### 4. Comments Suggestions And Typos Type: list of points. Other improvements that are not part of the concerns above, including suggested additional experiments, presentation fixes and typos. ### 5. Confidence Type: score, 1–5. How confident you are in your evaluation. - **5**: Positive that my evaluation is correct. I read the paper very carefully and am familiar with related work. - **4**: Quite sure. I tried to check the important points carefully. It's unlikely, though conceivable, that I missed something that should affect my ratings. - **3**: Pretty sure, but there's a chance I missed something. Although I have a good feel for this area in general, I did not carefully check the paper's details, e.g., the math or experimental design. - **2**: Willing to defend my evaluation, but it is fairly likely that I missed some details, didn't understand some central points, or can't be sure about the novelty of the work. - **1**: Not my area, or paper is very hard to understand. My evaluation is just an educated guess. ### 6. Soundness Type: score, 1–5 (step 0.5). Objective: does the study support its claims? Must be justified in the review text. - **5**: Excellent: This study is one of the most thorough I have seen, given its type. - **4**: Strong: This study provides sufficient support for all of its claims. Some extra experiments could be nice, but not essential. - **3**: Acceptable: This study provides sufficient support for its main claims. Some minor points may need extra support or details. - **2**: Poor: Some of the main claims are not sufficiently supported. There are major technical/methodological problems. - **1**: Major Issues: This study is not yet sufficiently thorough to warrant publication or is not relevant to ACL. ### 7. Excitement Type: score, 1–5 (step 0.5). Explicitly subjective: how exciting you personally find the work. - **5**: Highly Exciting: I would recommend this paper to others and/or attend its presentation in a conference. - **4**: Exciting: I would mention this paper to others and/or make an effort to attend its presentation in a conference. - **3**: Interesting: I might mention some points of this paper to others and/or attend its presentation in a conference if there's time. - **2**: Potentially Interesting: this paper does not resonate with me, but it might with others in the *ACL community. - **1**: Not Exciting: this paper does not resonate with me, and I don't think it would with others in the *ACL community. ### 8. Overall Assessment Type: score, 1–5 (step 0.5). An outcome recommendation (Award / Conference / Findings / Resubmit), not a quality score. - **5**: Consider for Award: I think this paper could be considered for an outstanding paper award at an *ACL conference (up to top 2.5% papers). - **4.5**: Borderline Award - **4**: Conference: I think this paper could be accepted to an *ACL conference. - **3.5**: Borderline Conference - **3**: Findings: I think this paper could be accepted to the Findings of the ACL. - **2.5**: Borderline Findings - **2**: Resubmit next cycle: I think this paper needs substantial revisions that can be completed by the next ARR cycle. - **1.5**: Resubmit after next cycle: I think this paper needs substantial revisions that cannot be completed by the next ARR cycle. - **1**: Do not resubmit: This paper has to be fully redone, or it is not relevant to the *ACL community (e.g., it is in no way related to computational processing of language). ### 9. Best Paper Justification (optional) Type: free text. Only if Overall Assessment is 4.5 or 5; otherwise write "N/A". ### 10. Limitations And Societal Impact Type: free text. Have the authors adequately discussed the limitations and potential positive and negative societal impacts of their work? ### 11. Ethical Concerns Type: free text. Concerns with respect to the ACL Code of Ethics and the Responsible NLP checklist. "None" if there are none. ### 12. Needs Ethics Review Type: single choice. Should the paper go to the ethics committee? Answer No or Yes (with a reason). - No - Yes ### 13. Reproducibility Type: score, 1–5. Could others reproduce the results? - **5**: They could easily reproduce the results. - **4**: They could mostly reproduce the results, but there may be some variation because of sample variance or minor variations in their interpretation of the protocol or method. - **3**: They could reproduce the results with some difficulty. The settings of parameters are underspecified or subjectively determined, and/or the training/evaluation data are not widely available. - **2**: They would be hard pressed to reproduce the results. The contribution depends on data that are simply not available outside the author's institution or consortium and/or not enough details are provided. - **1**: They would not be able to reproduce the results here no matter how hard they tried. ### 14. Datasets Type: score, 1–5. Value of any newly released datasets (1 if none). - **5**: Enabling: The newly released datasets should affect other people's choice of research or development projects to undertake. - **4**: Useful: I would recommend the new datasets to other researchers or developers for their ongoing work. - **3**: Potentially useful: Someone might find the new datasets useful for their work. - **2**: Documentary: The new datasets will be useful to study or replicate the reported research, although for other purposes they may have limited interest or limited usability. (Still a positive rating) - **1**: No usable datasets submitted. ### 15. Software Type: score, 1–5. Value of any newly released software (1 if none). - **5**: Enabling: The newly released software should affect other people's choice of research or development projects to undertake. - **4**: Useful: I would recommend the new software to other researchers or developers for their ongoing work. - **3**: Potentially useful: Someone might find the new software useful for their work. - **2**: Documentary: The new software will be useful to study or replicate the reported research, although for other purposes they may have limited interest or limited usability. (Still a positive rating) - **1**: No usable software released. ## Area chair meta-review ARR meta-reviews use the reviewers' Overall Assessment scale. Weigh arguments, not averages. Recommendation options: Conference · Findings · Resubmit next cycle · Resubmit after next cycle · Do not resubmit ## Responsible NLP Research Checklist 1. A1. Limitations section describing the work's limitations (required; missing it is grounds for desk rejection) 2. A2. Potential risks of the work discussed 3. B1. Creators of artifacts used are cited 4. B2. License and terms of use of artifacts discussed 5. B3. Use of existing artifacts consistent with their intended use 6. B4. Data checked for personally identifying or offensive content and anonymized 7. B5. Documentation of artifacts (domains, languages, linguistic phenomena, demographics) 8. B6. Relevant statistics (number of examples, train/test/dev splits) reported 9. C1. Number of parameters, total computational budget and computing infrastructure reported 10. C2. Experimental setup, hyperparameter search and best-found values discussed 11. C3. Descriptive statistics (e.g. error bars) reported, and whether results are single runs or aggregates 12. C4. Implementation, model and parameter settings of existing packages reported 13. D1. Full text of instructions given to human participants/annotators 14. D2. Recruitment and payment of participants, and whether payment was adequate 15. D3. Consent obtained from participants whose data is used 16. D4. Data collection protocol approved by an ethics review board 17. D5. Basic demographic and geographic characteristics of annotators reported 18. E1. Use of AI assistants (in research or writing) disclosed ## LLM policy ARR forbids using generative AI to create the first draft of a review or meta-review. Light language editing is allowed, but manuscripts and reviews must not be uploaded to non-privacy-preserving tools. Official policy: https://aclrollingreview.org/reviewerguidelines --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=eacl (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # AACL 2026 review form > Simulate AACL 2026 reviews in the ACL Rolling Review form: Paper Summary, Strengths, numbered Weaknesses, Soundness and Excitement (1–5 in half points), and the Overall Assessment that routes papers to AACL main, Findings or resubmission. - Venue: AACL-IJCNLP (Asia-Pacific Chapter of the ACL and International Joint Conference on NLP) - Form status: official (from the venue's public reviewer form) - Official source: https://aclrollingreview.org/reviewform - Overall score: Overall Assessment, 1–5 - Rough acceptance bar (our simulation): mean ≥ 3.5 - Web page: https://reviewpaper.ai/venues/aacl ## How reviewers read the paper AACL 2026 does not run its own reviewing: papers are reviewed in ACL Rolling Review and then committed to AACL, whose program chairs decide from the ARR reviews and meta-review. AACL-IJCNLP 2026 (Hengqin, China, November 2026) takes papers from the ARR May 2026 cycle (shared with EMNLP) or earlier; commitment ran in August 2026. ARR reviews are about half a page. The form separates objective Soundness (must be justified in the text) from subjective Excitement, and the Overall Assessment is an outcome recommendation (Award / Conference / Findings / Resubmit), all on 1–5 with half points. Weaknesses are a numbered list specific enough for an AC to follow; suggested extra work goes under Comments. The guidelines list lazy-reviewing heuristics to avoid, e.g. "not novel" without a reference, "not SOTA", "simple method", "only tested on English", "I would have done X", or listing limitations the authors already acknowledge. - Number the weaknesses; each must be concrete enough for an AC to follow without reading the paper. - Avoid ARR lazy-reviewing heuristics: no "not novel" without citing the prior work, no "not SOTA" as a reason to reject, no "the method is too simple", no "only tested on English" unless the claims are general, no "I would have done X", no listing limitations the authors already acknowledge as weaknesses. - Put suggested additional experiments under Comments, not Weaknesses. - Justify the Soundness score in the text; Excitement is explicitly personal taste. **Where the bar sits.** AACL-IJCNLP published a Findings volume in 2022 and 2023. In the Overall Assessment, 4 ("Conference") argues for the main conference and 3 ("Findings") for the Findings of the ACL volume, which the program chairs may offer to committed papers that are sound but less exciting. ARR publishes no cutoff; venue senior ACs decide. Rule of thumb: average Overall Assessment ≥ 3.5 → main conference plausible, 3.0–3.5 → Findings, < 3.0 → likely rejected. Main conference acceptance is about 20–25%, plus ~13–15% Findings. ## Reviewer form, field by field ### 1. Paper Summary Type: free text. Describe what the paper is about so the AC and action editor can catch any misunderstanding. ### 2. Summary Of Strengths Type: list of points. The major reasons to publish this paper at a selective *ACL venue. ### 3. Summary Of Weaknesses Type: list of points. Numbered list of the concerns that would make you prioritize other high-quality papers. Each must be specific enough for an AC to follow without reading the paper. Do NOT include suggested extra work here (that goes in Comments), and do not list limitations the authors already acknowledge. ### 4. Comments Suggestions And Typos Type: list of points. Other improvements that are not part of the concerns above, including suggested additional experiments, presentation fixes and typos. ### 5. Confidence Type: score, 1–5. How confident you are in your evaluation. - **5**: Positive that my evaluation is correct. I read the paper very carefully and am familiar with related work. - **4**: Quite sure. I tried to check the important points carefully. It's unlikely, though conceivable, that I missed something that should affect my ratings. - **3**: Pretty sure, but there's a chance I missed something. Although I have a good feel for this area in general, I did not carefully check the paper's details, e.g., the math or experimental design. - **2**: Willing to defend my evaluation, but it is fairly likely that I missed some details, didn't understand some central points, or can't be sure about the novelty of the work. - **1**: Not my area, or paper is very hard to understand. My evaluation is just an educated guess. ### 6. Soundness Type: score, 1–5 (step 0.5). Objective: does the study support its claims? Must be justified in the review text. - **5**: Excellent: This study is one of the most thorough I have seen, given its type. - **4**: Strong: This study provides sufficient support for all of its claims. Some extra experiments could be nice, but not essential. - **3**: Acceptable: This study provides sufficient support for its main claims. Some minor points may need extra support or details. - **2**: Poor: Some of the main claims are not sufficiently supported. There are major technical/methodological problems. - **1**: Major Issues: This study is not yet sufficiently thorough to warrant publication or is not relevant to ACL. ### 7. Excitement Type: score, 1–5 (step 0.5). Explicitly subjective: how exciting you personally find the work. - **5**: Highly Exciting: I would recommend this paper to others and/or attend its presentation in a conference. - **4**: Exciting: I would mention this paper to others and/or make an effort to attend its presentation in a conference. - **3**: Interesting: I might mention some points of this paper to others and/or attend its presentation in a conference if there's time. - **2**: Potentially Interesting: this paper does not resonate with me, but it might with others in the *ACL community. - **1**: Not Exciting: this paper does not resonate with me, and I don't think it would with others in the *ACL community. ### 8. Overall Assessment Type: score, 1–5 (step 0.5). An outcome recommendation (Award / Conference / Findings / Resubmit), not a quality score. - **5**: Consider for Award: I think this paper could be considered for an outstanding paper award at an *ACL conference (up to top 2.5% papers). - **4.5**: Borderline Award - **4**: Conference: I think this paper could be accepted to an *ACL conference. - **3.5**: Borderline Conference - **3**: Findings: I think this paper could be accepted to the Findings of the ACL. - **2.5**: Borderline Findings - **2**: Resubmit next cycle: I think this paper needs substantial revisions that can be completed by the next ARR cycle. - **1.5**: Resubmit after next cycle: I think this paper needs substantial revisions that cannot be completed by the next ARR cycle. - **1**: Do not resubmit: This paper has to be fully redone, or it is not relevant to the *ACL community (e.g., it is in no way related to computational processing of language). ### 9. Best Paper Justification (optional) Type: free text. Only if Overall Assessment is 4.5 or 5; otherwise write "N/A". ### 10. Limitations And Societal Impact Type: free text. Have the authors adequately discussed the limitations and potential positive and negative societal impacts of their work? ### 11. Ethical Concerns Type: free text. Concerns with respect to the ACL Code of Ethics and the Responsible NLP checklist. "None" if there are none. ### 12. Needs Ethics Review Type: single choice. Should the paper go to the ethics committee? Answer No or Yes (with a reason). - No - Yes ### 13. Reproducibility Type: score, 1–5. Could others reproduce the results? - **5**: They could easily reproduce the results. - **4**: They could mostly reproduce the results, but there may be some variation because of sample variance or minor variations in their interpretation of the protocol or method. - **3**: They could reproduce the results with some difficulty. The settings of parameters are underspecified or subjectively determined, and/or the training/evaluation data are not widely available. - **2**: They would be hard pressed to reproduce the results. The contribution depends on data that are simply not available outside the author's institution or consortium and/or not enough details are provided. - **1**: They would not be able to reproduce the results here no matter how hard they tried. ### 14. Datasets Type: score, 1–5. Value of any newly released datasets (1 if none). - **5**: Enabling: The newly released datasets should affect other people's choice of research or development projects to undertake. - **4**: Useful: I would recommend the new datasets to other researchers or developers for their ongoing work. - **3**: Potentially useful: Someone might find the new datasets useful for their work. - **2**: Documentary: The new datasets will be useful to study or replicate the reported research, although for other purposes they may have limited interest or limited usability. (Still a positive rating) - **1**: No usable datasets submitted. ### 15. Software Type: score, 1–5. Value of any newly released software (1 if none). - **5**: Enabling: The newly released software should affect other people's choice of research or development projects to undertake. - **4**: Useful: I would recommend the new software to other researchers or developers for their ongoing work. - **3**: Potentially useful: Someone might find the new software useful for their work. - **2**: Documentary: The new software will be useful to study or replicate the reported research, although for other purposes they may have limited interest or limited usability. (Still a positive rating) - **1**: No usable software released. ## Area chair meta-review ARR meta-reviews use the reviewers' Overall Assessment scale. Weigh arguments, not averages. Recommendation options: Conference · Findings · Resubmit next cycle · Resubmit after next cycle · Do not resubmit ## Responsible NLP Research Checklist 1. A1. Limitations section describing the work's limitations (required; missing it is grounds for desk rejection) 2. A2. Potential risks of the work discussed 3. B1. Creators of artifacts used are cited 4. B2. License and terms of use of artifacts discussed 5. B3. Use of existing artifacts consistent with their intended use 6. B4. Data checked for personally identifying or offensive content and anonymized 7. B5. Documentation of artifacts (domains, languages, linguistic phenomena, demographics) 8. B6. Relevant statistics (number of examples, train/test/dev splits) reported 9. C1. Number of parameters, total computational budget and computing infrastructure reported 10. C2. Experimental setup, hyperparameter search and best-found values discussed 11. C3. Descriptive statistics (e.g. error bars) reported, and whether results are single runs or aggregates 12. C4. Implementation, model and parameter settings of existing packages reported 13. D1. Full text of instructions given to human participants/annotators 14. D2. Recruitment and payment of participants, and whether payment was adequate 15. D3. Consent obtained from participants whose data is used 16. D4. Data collection protocol approved by an ethics review board 17. D5. Basic demographic and geographic characteristics of annotators reported 18. E1. Use of AI assistants (in research or writing) disclosed ## LLM policy ARR forbids using generative AI to create the first draft of a review or meta-review. Light language editing is allowed, but manuscripts and reviews must not be uploaded to non-privacy-preserving tools. Official policy: https://aclrollingreview.org/reviewerguidelines --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=aacl (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # COLM 2026 review form > Simulate COLM 2026-style reviews: Summary, Reasons To Accept, Reasons To Reject, Questions, assessed across COLM's five dimensions (empiricism, technological impact, ambition, understanding, clarity and honesty), with a 1–10 rating and confidence. - Venue: Conference on Language Modeling - Form status: modeled, not official. COLM's OpenReview form is not public. The five evaluation dimensions come from the official 2026 review guidelines; the field list (Summary, Reasons To Accept/Reject, Questions) and the 1–10 rating and 1–5 confidence wording follow earlier COLM reviews and OpenReview defaults, not a confirmed 2026 form. - Official source: https://colm.cc/Conferences/2026/ReviewGuidelines - Overall score: Rating, 1–10 - Rough acceptance bar (our simulation): mean ≥ 6 - Web page: https://reviewpaper.ai/venues/colm ## How reviewers read the paper COLM asks reviewers to weigh five dimensions without mechanistic measures of quality: Empiricism, Data and Evaluation; Technological Impact; Ambition, Vision, Forward-outlook; Understanding Depth, Principled Approach; and Clarity, Honesty and Trust. Weakness on one dimension is not an automatic reject, especially for ambitious and honest work, while strength on one does not compensate for being very weak on others. Small-scale work should not be penalized for lacking large compute. Reviews of accepted papers become public. - Do not apply mechanistic quality measures (e.g. "no SOTA", "too few benchmarks") as automatic rejection reasons. - Do not penalize small-scale or academic-compute work for lacking large-scale experiments. - Credit honesty about limitations; do not treat an honest limitations section as a weakness. - Respectful tone; reviews of accepted papers are published. **Where the bar sits.** COLM accepted 28.9% in 2024. Accepted papers averaged about 6.4–6.5 on the 1–10 rating in 2024 and 2025 (Paper Copilot). Rough rule: mean ≥ 6 competitive, 5.25–6 borderline. ## Reviewer form, field by field ### 1. Summary Type: free text. What the paper does and claims, in your own words. ### 2. Reasons To Accept Type: list of points. Strengths across COLM's dimensions: Empiricism, Data and Evaluation; Technological Impact; Ambition, Vision, Forward-outlook; Understanding Depth, Principled Approach; Clarity, Honesty and Trust. ### 3. Reasons To Reject Type: list of points. Weaknesses across the same dimensions. Weakness on one dimension is not by itself a reason to reject, especially for ambitious work honest about its limits; but strength on one dimension does not offset being very weak on others. Do not penalize small-scale work for lacking large compute. ### 4. Questions To Authors Type: list of points. Questions for the rebuttal whose answers could change your assessment. ### 5. Rating Type: score, 1–10. Overall rating. - **10**: Top 5% of accepted papers, seminal paper - **9**: Top 15% of accepted papers, strong accept - **8**: Top 50% of accepted papers, clear accept - **7**: Good paper, accept - **6**: Marginally above acceptance threshold - **5**: Marginally below acceptance threshold - **4**: Ok but not good enough - rejection - **3**: Clear rejection - **2**: Strong rejection - **1**: Trivial or wrong ### 6. Confidence Type: score, 1–5. How certain you are about your evaluation. - **5**: The reviewer is absolutely certain that the evaluation is correct and very familiar with the relevant literature - **4**: The reviewer is confident but not absolutely certain that the evaluation is correct - **3**: The reviewer is fairly confident that the evaluation is correct - **2**: The reviewer is willing to defend the evaluation, but it is quite likely that the reviewer did not understand central parts of the paper - **1**: The reviewer's evaluation is an educated guess ### 7. Ethics Concerns Type: free text. "None", or the ethical concern that should be flagged for ethics review. ## Area chair meta-review COLM ACs make the initial decision, may dismiss clearly low-quality reviews, and discuss borderline cases with the PCs. Recommendation options: Accept · Reject ## COLM author statements 1. Optional ethics statement (up to one page, outside the page limit) 2. Optional reproducibility statement (up to one page) 3. LLM use that originated ideas, wrote content (incl. references), generated data or plots, or did evaluation disclosed 4. Anonymized: no acknowledgments or identifying links ## LLM policy COLM 2026 requires reviewers to disclose any LLM use, cosmetic or substantive; reviewers remain fully responsible for their reviews. Official policy: https://colm.cc/Conferences/2026/CallForPapers --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=colm (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # CVPR 2026 review form > Simulate CVPR 2026-style reviews: Summary, Strengths, Weaknesses, a five-level Preliminary Recommendation (Strong Accept to Reject) and a justification with points for the rebuttal, written in CVPR's depersonalized voice. - Venue: IEEE/CVF Conference on Computer Vision and Pattern Recognition - Form status: modeled, not official. The CVPR review form is not public. Field order is reconstructed from the 2024–2025 OpenReview form; the five recommendation labels come from the 2026 reviewer training material, and the 1–5 numeric mapping is ours. - Official source: https://cvpr.thecvf.com/Conferences/2026/ReviewerTrainingMaterial - Overall score: Preliminary Recommendation, 1–5 - Rough acceptance bar (our simulation): mean ≥ 3.5 - Web page: https://reviewpaper.ai/venues/cvpr ## How reviewers read the paper CVPR's training material asks for a summary in the reviewer's own words, specific evidence-based strengths and weaknesses (novelty concerns must cite the specific prior work), and a recommendation with a justification that tells the authors what the one-page rebuttal should address. Do not reject solely for missing SOTA, missing comparisons to arXiv-only work, easily fixable errors, or a missing limitations section. Refer to "the paper", not "you" or "the authors". - Refer to "the paper" — never "you" or "the authors". - Professional, courteous tone with concrete suggestions. - Short or generic reviews are flagged as irresponsible; each weakness needs specifics. - Do not request new experiments as a condition for the rebuttal (the rebuttal cannot add new experiments or contributions). **Where the bar sits.** CVPR 2026 accepted about 25% of 16,092 submissions. No score statistics are public; on this 1–5 mapping, a mean of ≥ 3.5 (mostly Weak Accept or better) is a rough proxy for acceptance. ## Reviewer form, field by field ### 1. Summary Type: free text. 2–4 sentences in your own words: what the paper does and claims. ### 2. Strengths Type: list of points. Specific and evidence-based. Avoid generic praise such as "the paper is interesting". ### 3. Weaknesses Type: list of points. Specific and detailed — bullet lists of short phrases are not a detailed review. A novelty concern must cite the specific prior work and explain how closely it relates. Do not reject solely for not beating SOTA, for lacking comparisons to arXiv-only work, or for easily fixable errors. ### 4. Preliminary Recommendation Type: score, 1–5. Strong Accept: significant contribution with strong evidence and broad interest. Weak Accept: sound and well motivated but incremental or not best on every benchmark. Borderline: reasonable but with gaps in evaluation or positioning. Weak Reject: limited novelty relative to closely related prior work. Reject: fundamental technical issues. - **5**: Strong Accept - **4**: Weak Accept - **3**: Borderline - **2**: Weak Reject - **1**: Reject ### 5. Justification For Recommendation And Suggestions For Rebuttal Type: free text. Connect the rating to the assessment above, and state precisely what the rebuttal should address. ### 6. Confidence Level Type: score, 1–5. How certain you are about your assessment. - **5**: You are absolutely certain about your assessment. You are very familiar with the related work and checked the math/other details carefully. - **4**: You are confident in your assessment, but not absolutely certain. It is unlikely, but not impossible, that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. - **3**: You are fairly confident in your assessment. It is possible that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked. - **2**: You are willing to defend your assessment, but it is quite likely that you did not understand the central parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked. - **1**: Your assessment is an educated guess. The submission is not in your area or the submission was difficult to understand. Math/other details were not carefully checked. ### 7. Dataset Attribution Type: free text. Are all data assets properly cited? Flag any use of withdrawn or deprecated datasets. "No concerns" otherwise. ### 8. Ethics Flag Type: free text. "No" or a brief description of the ethical concern (e.g. human-subjects data without IRB statement). ## Area chair meta-review CVPR ACs work in triplets; the meta-review must justify the recommendation from the reviews and discussion. Recommendation options: Award candidate · Oral · Highlight · Poster · Findings · Reject ## CVPR author requirements 1. Main text within 8 pages plus references 2. Anonymized submission 3. Negative societal impact discussed where relevant 4. IRB approval stated for human-subjects data 5. All datasets cited; no withdrawn datasets used 6. No hallucinated or unverifiable citations ## LLM policy CVPR 2026 does not allow LLMs to be used to write reviews or meta-reviews, and no substantial paper or review content may be shared with an LLM. Violations can lead to desk rejection of the reviewer's own papers. Official policy: https://cvpr.thecvf.com/Conferences/2026/ReviewerGuidelines --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=cvpr (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # ICCV 2025 review form > Simulate ICCV-style reviews: Summary, Strengths, Weaknesses, a six-level preliminary recommendation, a justification with points for the rebuttal, and confidence, in the CVF depersonalized voice. - Venue: IEEE/CVF International Conference on Computer Vision - Form status: modeled, not official. Modeled on ICCV 2025 (ICCV 2027 guidelines are not out yet). The form is not public: field order follows the CVF/CVPR 2025 form, and the six-level recommendation scale is reconstructed from a community post, not official wording. - Official source: https://iccv.thecvf.com/Conferences/2025/ReviewerGuidelines - Overall score: Preliminary Recommendation, 1–6 - Rough acceptance bar (our simulation): mean ≥ 4 - Web page: https://reviewpaper.ai/venues/iccv ## How reviewers read the paper ICCV reviewers argue strengths and weaknesses rather than listing them, give a preliminary recommendation and a justification that tells the authors what the one-page rebuttal should address, then update their rating and justification after the rebuttal. Comparisons should only be requested to work public sufficiently ahead of the deadline; not beating SOTA is not on its own grounds for rejection; no substantial new experiments in the rebuttal. - Refer to "the paper" — never "you" or "the authors". - No belittling or sarcastic remarks; professional, courteous tone with concrete suggestions. - Back up any "done before" claim with specific references and explain how close the prior work is. - Not beating state of the art, or not citing/beating an arXiv-only paper, is not on its own grounds for rejection. - Do not request substantial new experiments for the rebuttal. **Where the bar sits.** ICCV 2025 accepted 2,701 of 11,239 valid submissions (24%, "not pre-determined"). No score statistics are public; on this modeled 1–6 scale a mean of ≥ 4 (Borderline Accept) is a rough proxy for acceptance. ## Reviewer form, field by field ### 1. Summary Type: free text. The key ideas, contributions and their significance, in your own words. ### 2. Strengths Type: list of points. What gives the paper value, each point argued with specifics; short bullet fragments are not acceptable. ### 3. Weaknesses Type: list of points. What detracts from the contributions, each argued with specifics. A novelty concern must cite the specific prior work. Do not reject solely for missing SOTA or missing comparisons to work not public well before the deadline. ### 4. Preliminary Recommendation Type: score, 1–6. Rating before the rebuttal; reviewers update it with a final justification afterwards. - **6**: Accept - **5**: Weak Accept - **4**: Borderline Accept - **3**: Borderline Reject - **2**: Weak Reject - **1**: Reject ### 5. Justification For Recommendation And Suggestions For Rebuttal Type: free text. Connect the rating to the assessment above and state precisely what the one-page rebuttal should address. ### 6. Confidence Type: score, 1–5. How certain you are about your assessment. - **5**: You are absolutely certain about your assessment. You are very familiar with the related work and checked the math/other details carefully. - **4**: You are confident in your assessment, but not absolutely certain. It is unlikely, but not impossible, that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. - **3**: You are fairly confident in your assessment. It is possible that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked. - **2**: You are willing to defend your assessment, but it is quite likely that you did not understand the central parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked. - **1**: Your assessment is an educated guess. The submission is not in your area or the submission was difficult to understand. Math/other details were not carefully checked. ### 7. Additional Comments (optional) Type: list of points. Minor issues, typos and suggestions that do not affect the recommendation. ## Area chair meta-review ICCV ACs work in triplets, decide accept/reject and nominate highlights, orals and awards. Recommendation options: Award candidate · Oral · Highlight · Poster · Reject ## ICCV author requirements 1. Main text within 8 pages plus references 2. Anonymized submission; complete OpenReview profiles for all authors 3. At most 25 submissions per author 4. No citations to non-existent material ## LLM policy ICCV 2025 does not allow LLMs or online chatbots in any part of the reviewing process; LLM-generated reviews are "highly irresponsible" and can lead to desk rejection of all the reviewer's papers. Official policy: https://iccv.thecvf.com/Conferences/2025/ReviewerGuidelines --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=iccv (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # ECCV 2026 review form > Simulate ECCV 2026-style reviews: Summary, Strengths and Weaknesses judged against the paper's contribution type, a rating with justification, and additional comments, in the CVF/ECVA depersonalized voice. - Venue: European Conference on Computer Vision - Form status: modeled, not official. ECCV's OpenReview form is not public. The sections (Summary, Strengths, Weaknesses, Rating and Justification, Additional Comments) and contribution types come from the official 2026 reviewer tutorial slides; the numeric rating and confidence scales are modeled on CVPR, not official. - Official source: https://eccv.ecva.net/Conferences/2026/ReviewerGuide - Overall score: Rating, 1–5 - Rough acceptance bar (our simulation): mean ≥ 3.5 - Web page: https://reviewpaper.ai/venues/eccv ## How reviewers read the paper ECCV 2026 reviewers write a summary ("your abstract of the paper"), substantive strengths and weaknesses judged against the paper's declared contribution type (short bullet lists are not acceptable), a rating with a justification that weighs the key strengths and weaknesses, and minor additional comments; after the rebuttal they give a final recommendation and justification. Avoid the novelty fallacy, pure opinions, blank assertions, and demands to beat arXiv papers. - Refer to "the paper" — never "you" or "the authors". - No belittling or sarcastic remarks; professional, courteous tone with concrete suggestions. - Back up any "done before" claim with specific references and explain how close the prior work is. - Not beating state of the art, or not citing/beating an arXiv-only paper, is not on its own grounds for rejection. - Do not request substantial new experiments for the rebuttal. - Avoid the novelty fallacy, blank assertions and pure opinions; strengths and weaknesses must be argued, not listed. **Where the bar sits.** ECCV 2026 accepted about 27.5% of ~10,500 submissions (secondary sources; ECCV 2024: 27.9%). No score statistics are public; on this modeled 1–5 scale a mean of ≥ 3.5 is a rough proxy for acceptance. ## Reviewer form, field by field ### 1. Contribution Type Type: single choice. The contribution type the authors most plausibly declared. Strengths and weaknesses are judged against it (e.g. a Datasets/Benchmarks paper is not expected to propose a novel method); any departure must be explained in the justification. - Algorithms/General - Theory/Foundational - Applied/Systems - Datasets/Benchmarks - Concept & Feasibility ### 2. Summary Type: free text. The key ideas, contributions and their significance, in your own words. ### 3. Strengths Type: list of points. What gives the paper value, each point argued with specifics; short bullet fragments are not acceptable. Judge value against the declared contribution type. ### 4. Weaknesses Type: list of points. What detracts from the contributions, each argued with specifics. A novelty concern must cite the specific prior work. Judge against the declared contribution type. ### 5. Rating Type: score, 1–5. Preliminary rating before the rebuttal. - **5**: Strong Accept - **4**: Weak Accept - **3**: Borderline - **2**: Weak Reject - **1**: Reject ### 6. Justification Of Rating Type: free text. Why the paper should or should not be accepted: which strengths and weaknesses matter most and how they weigh against each other given the contribution type, and what the rebuttal should address. ### 7. Additional Comments Type: list of points. Minor suggestions, questions and corrections that are not crucial to the recommendation. ### 8. Confidence Type: score, 1–5. How certain you are about your assessment. - **5**: You are absolutely certain about your assessment. You are very familiar with the related work and checked the math/other details carefully. - **4**: You are confident in your assessment, but not absolutely certain. It is unlikely, but not impossible, that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. - **3**: You are fairly confident in your assessment. It is possible that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked. - **2**: You are willing to defend your assessment, but it is quite likely that you did not understand the central parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked. - **1**: Your assessment is an educated guess. The submission is not in your area or the submission was difficult to understand. Math/other details were not carefully checked. ## Area chair meta-review ECCV ACs decide accept/reject and nominate orals and award candidates; accepting against unanimous reviewers needs exceptional justification. Do not lean on reviewer confidence scores. Recommendation options: Award candidate · Oral · Poster · Reject ## ECCV author requirements 1. Main paper at most 14 pages excluding references 2. Contribution type declared (Algorithms/General, Theory/Foundational, Applied/Systems, Datasets/Benchmarks, Concept & Feasibility) 3. Anonymized; no external links that expand content, and no prompt injection 4. Datasets/Benchmarks papers: data public by camera-ready 5. No citations to non-existent material ## LLM policy ECCV 2026 does not allow LLMs to write reviews or meta-reviews, locally or via API, and no substantial paper or review content may be shared with an LLM (including for translation). Grammar checkers and background research are allowed. Violations can desk-reject the reviewer's own papers. Official policy: https://eccv.ecva.net/Conferences/2026/ReviewerGuide --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=eccv (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # WACV 2027 review form > Simulate WACV 2027-style reviews for the Algorithms, Applications or Evaluations & Datasets track: Summary, Strengths, Weaknesses, a five-level rating with justification, and confidence; round-1 papers can get Revise and Resubmit. - Venue: IEEE/CVF Winter Conference on Applications of Computer Vision - Form status: modeled, not official. WACV's OpenReview form is not public. Fields follow the CVF pattern and the tracks and round-1 outcomes come from the official 2027 guides; the five-level rating is modeled from labels used in WACV 2026 AC templates (Strong Reject / Reject / Neutral / Accept), not official wording. - Official source: https://wacv.thecvf.com/Conferences/2027/ReviewerGuidelines - Overall score: Rating, 1–5 - Rough acceptance bar (our simulation): mean ≥ 3.5 - Web page: https://reviewpaper.ai/venues/wacv ## How reviewers read the paper WACV runs two rounds. Round-1 papers get Accept, Revise and Resubmit (a revised paper plus one-page rebuttal returns to the same reviewers about two weeks later) or Reject; round-2 papers are accepted or rejected with no rebuttal. Reviewers argue strengths and weaknesses in detail and judge Applications papers on systems-level innovation, domain novelty and comparative assessment rather than algorithmic novelty alone. - Refer to "the paper" — never "you" or "the authors". - No belittling or sarcastic remarks; professional, courteous tone with concrete suggestions. - Back up any "done before" claim with specific references and explain how close the prior work is. - Not beating state of the art, or not citing/beating an arXiv-only paper, is not on its own grounds for rejection. - Do not request substantial new experiments for the rebuttal. - Use neutral references ("the paper", "R1"); do not ask the paper to cite your own work. **Where the bar sits.** WACV 2026 accepted 832 of 2,550 submissions (33%): round 1 38% (84 direct accepts, 439 of 736 revisions accepted), round 2 26%. No score statistics are public; on this modeled 1–5 scale a mean of ≥ 3.5 is a rough proxy for acceptance or a revise-and-resubmit invitation. ## Reviewer form, field by field ### 1. Track Type: single choice. The track the paper was most plausibly submitted to. Applications papers are judged on systems-level innovation, novelty of the application domain and comparative assessment, not solely on algorithmic novelty; Evaluations & Datasets papers are judged like the NeurIPS E&D track. - Algorithms - Applications - Evaluations & Datasets ### 2. Summary Type: free text. The key ideas, contributions and their significance, in your own words. ### 3. Strengths Type: list of points. What gives the paper value, each point argued with specifics; short bullet fragments are not acceptable. A single short sentence or a few words do not suffice. ### 4. Weaknesses Type: list of points. What detracts from the contributions, each argued with specifics. A novelty concern must cite the specific prior work. A single short sentence or a few words do not suffice. ### 5. Rating Type: score, 1–5. Overall rating for the track's criteria. - **5**: Strong Accept - **4**: Accept - **3**: Neutral - **2**: Reject - **1**: Strong Reject ### 6. Justification And Requested Revisions Type: free text. Connect the rating to the assessment. For round-1 papers, list precisely the revisions that a two-week Revise-and-Resubmit could address, and which issues (if any) could not be fixed in that window. ### 7. Confidence Type: score, 1–5. How certain you are about your assessment. - **5**: You are absolutely certain about your assessment. You are very familiar with the related work and checked the math/other details carefully. - **4**: You are confident in your assessment, but not absolutely certain. It is unlikely, but not impossible, that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. - **3**: You are fairly confident in your assessment. It is possible that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked. - **2**: You are willing to defend your assessment, but it is quite likely that you did not understand the central parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked. - **1**: Your assessment is an educated guess. The submission is not in your area or the submission was difficult to understand. Math/other details were not carefully checked. ## Area chair meta-review WACV AC pairs decide; overruling all three reviewers is not acceptable, and departing from the reviewer majority must be raised with the PCs. Recommendation options: Accept · Revise and Resubmit · Reject ## WACV author requirements 1. Main text within 8 pages plus references; abstract at most 5,000 characters 2. Track chosen: Algorithms, Applications, or Evaluations & Datasets 3. Anonymized submission; supplementary material adds no results on new datasets or from an improved method 4. Reproducibility checklist (recommended) ## LLM policy WACV lists reviews generated by LLMs as highly irresponsible, which can lead to desk rejection of the reviewer's own papers. Official policy: https://wacv.thecvf.com/Conferences/2027/ReviewerGuidelines --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=wacv (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # AAAI 2026 review form > Simulate AAAI-style reviews: Summary, Strengths and Weaknesses, Questions for the Authors, four criterion ratings, and the 1–8 Overall Evaluation, written in AAAI's depersonalized voice. - Venue: AAAI Conference on Artificial Intelligence - Form status: modeled, not official. The AAAI review form is not public. The field list follows a published description of the AAAI-25 form; the four criterion labels (Excellent/Good/Fair/Poor) and the 1–8 overall labels are modeled on AAAI/IJCAI conventions, not official wording. - Official source: https://aaai.org/conference/aaai/aaai-26/instructions-for-aaai-26-reviewers/ - Overall score: Overall Evaluation, 1–8 - Rough acceptance bar (our simulation): mean ≥ 5 - Web page: https://reviewpaper.ai/venues/aaai ## How reviewers read the paper AAAI reviewers assess the paper's story (problem, limits of the state of the art, whether results support the claims, key novelty), presentation (soundness, clarity, unstated assumptions, reproducibility, limitations), related work, and evaluations (baselines, metrics, datasets, error analysis), separating essential from optional revisions. Reviews must be depersonalized and must not duplicate content from other reviews, including AAAI's own AI review. Review is two-phase: papers with clearly negative Phase-1 ratings are rejected before author feedback. - Depersonalize: avoid "I", "you" and "the authors"; write "the paper" or "this reviewer". - Professional and respectful. - Separate essential revisions from optional suggestions. **Where the bar sits.** AAAI-26 accepted 17.6% of 23,680 submissions, and Phase 1 removes many papers before author feedback. No public score statistics; rough rule: mean ≥ 5 (of 8) needed to be competitive. ## Reviewer form, field by field ### 1. Summary Type: free text. The main claims and contributions of the paper. ### 2. Strengths Type: list of points. Part of "Strengths And Weaknesses". Cover the story (problem, limits of the state of the art, key novel contribution), presentation, placement in related work, and evaluation. ### 3. Weaknesses Type: list of points. Part of "Strengths And Weaknesses". Do the results support the claims? Unstated assumptions, missing reproducibility detail, missing related work, weak baselines/metrics/datasets/error analysis. Separate essential revisions from optional ones. ### 4. Questions For The Authors Type: list of points. Questions whose answers could change the evaluation. These feed the single author-feedback window. ### 5. Significance Of The Problem Type: score, 1–4. Excellent / Good / Fair / Poor. - **4**: excellent - **3**: good - **2**: fair - **1**: poor ### 6. Justification Of Approach Type: score, 1–4. Excellent / Good / Fair / Poor. - **4**: excellent - **3**: good - **2**: fair - **1**: poor ### 7. Quality Of Evaluation Type: score, 1–4. Excellent / Good / Fair / Poor. - **4**: excellent - **3**: good - **2**: fair - **1**: poor ### 8. Reproducibility And Facilitation Of Follow Up Work Type: score, 1–4. Excellent / Good / Fair / Poor. - **4**: excellent - **3**: good - **2**: fair - **1**: poor ### 9. Ethical Considerations Type: free text. Responsible research practice: human subjects, sensitive data, bias. "None" if not applicable. ### 10. Overall Evaluation Type: score, 1–8. Overall recommendation on AAAI's eight-point scale. - **8**: Strong accept (top 5% of accepted papers) - **7**: Accept (top 15% of accepted papers) - **6**: Weak accept - **5**: Borderline accept - **4**: Borderline reject - **3**: Weak reject - **2**: Reject - **1**: Strong reject ### 11. Confidence Type: score, 1–5. How certain you are about your assessment. - **5**: You are absolutely certain about your assessment. You are very familiar with the related work and checked the math/other details carefully. - **4**: You are confident in your assessment, but not absolutely certain. It is unlikely, but not impossible, that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. - **3**: You are fairly confident in your assessment. It is possible that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked. - **2**: You are willing to defend your assessment, but it is quite likely that you did not understand the central parts of the submission or that you are unfamiliar with some pieces of related work. Math/other details were not carefully checked. - **1**: Your assessment is an educated guess. The submission is not in your area or the submission was difficult to understand. Math/other details were not carefully checked. ## Area chair meta-review AAAI SPCs write the meta-review and the AC confirms the decision. Recommendation options: Accept · Reject ## AAAI Reproducibility Checklist 1. General: motivation, contributions and pedagogical references clearly stated 2. Theory: all assumptions stated and complete proofs provided 3. Datasets: new datasets described and released (or reasons given); licensing of existing datasets 4. Code: source code for experiments released or will be released 5. Experiments: random seeds, hardware, hyperparameters and their selection reported 6. Experiments: evaluation metrics motivated; statistical significance / variance reported 7. Page limit: 7 pages of content plus up to 2 pages of references ## LLM policy No explicit AAAI reviewer-LLM rule was found, but reviewer confidentiality rules in practice forbid uploading submissions to external LLMs. AAAI supplies its own official AI review, and reviewers must not duplicate content from other reviews. Official policy: https://aaai.org/conference/aaai/aaai-26/ethical-guidelines-for-aaai-26-reviewers/ --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=aaai (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # IJCAI 2026 review form > Simulate IJCAI-ECAI 2026-style reviews: contribution summary, strengths, opportunities to improve, pressing rebuttal questions, the nine-level overall assessment with justification, reproducibility, expertise and confidence. - Venue: International Joint Conference on Artificial Intelligence - Form status: modeled, not official. IJCAI-ECAI 2026 did not publish its review form. Fields and labels follow the official IJCAI-ECAI 2022 form (the most recent public one); the 2026 pages confirm the criteria, the ethics flag and the Convincing / Credible / Irreproducible reproducibility levels. Numbering of the 9-level overall scale is ours. - Official source: https://ijcai-22.org/wp-content/uploads/2022/01/Review-form-IJCAI-ECAI-2022-2.pdf - Overall score: Overall Assessment, 1–9 - Rough acceptance bar (our simulation): mean ≥ 5 - Web page: https://reviewpaper.ai/venues/ijcai ## How reviewers read the paper IJCAI-ECAI 2026 reviews in two phases: two Phase-1 reviews decide summary rejection without author response, then Phase 2 adds reviews, a one-page author response (no new results) and an AC meta-review cross-checked by a senior AC. Criteria: novelty and originality, correctness, clarity, significance of results, potential impact and presentation quality. Questions for authors must be pressing and answerable without new experiments; missing code, data or full proofs should not be the main reason to reject. - Keep the summary free of evaluation. - Ask only pressing rebuttal questions that do not require new results. - Do not reject mainly for missing code, data or full proofs; rate reproducibility instead. - Professional and unbiased; reviewer identities stay confidential. **Where the bar sits.** IJCAI-ECAI 2026 accepted about 13.5% of main-track papers (secondary source; 2025: 19.3%). A summary-reject phase removes papers whose two Phase-1 reviews are not positive enough, and a rebuttal rarely rescues reject-leaning reviews. On the 1–9 scale a mean of ≥ 5 (Borderline Accept) is needed to be competitive; most accepts average 6+. ## Reviewer form, field by field ### 1. Summary Of Main Contributions Type: free text. Briefly summarize the main contributions in your own words. Do not include your evaluation here. ### 2. Main Strengths Type: list of points. Focus on novelty, soundness, significance and expected impact, relevance to AI, clarity of exposition, and reproducibility (credibility). ### 3. Opportunities To Improve The Paper Type: list of points. Weaknesses and how the paper could be improved, most important first. ### 4. Pressing Questions For The Authors In The Rebuttal Type: list of points. Only questions whose answers could directly change your evaluation and that do not require new results (the one-page rebuttal allows no new experiments). ### 5. Overall Assessment Type: score, 1–9. Overall assessment. Borderline levels sparingly. - **9**: Award Quality: Truly groundbreaking work. This paper is likely to change our field. It should be considered for a best paper award. (Top 5% of accepted IJCAI papers.) - **8**: Strong Accept. Inspiring. An excellent paper. This paper changed my thinking on this topic. I will fight for acceptance. (Top 15% of accepted IJCAI papers). - **7**: Clear Accept. Interesting. A very good submission. I learned a lot from this paper. I vote and argue for acceptance. (Top 50% of accepted IJCAI papers.) - **6**: Weak Accept. Useful. A good paper. The results and insights will benefit the field. I believe it should be accepted. - **5**: Borderline Accept. Marginally above the acceptance threshold. Technically correct, but not particularly exciting or inspiring. Could be accepted more or less in its current form. Not a big loss if it is not included in the program. Please use sparingly. - **4**: Borderline Reject. Marginally below the acceptance threshold. The paper has merits but there are key weaknesses. It would benefit from another revision. Can be rejected. But having it in the program would not be that bad. Please use sparingly. - **3**: Weak Reject. Not good enough. Limited contribution. Needs more work and/or revisions. I believe it should be rejected. - **2**: Clear Reject. Unclear contribution. Needs to be reworked and/or major revision. I vote and will argue for rejection. - **1**: Strong Reject. Contribution is flawed. I will strongly argue for rejection. ### 6. Justification Of Score Type: free text. Justify your score in a few lines. ### 7. Reproducibility Type: single choice. Convincing: results can be reproduced, possibly with some effort; key resources available (or good reasons given) and key details well described. Credible: resources unavailable, but details suffice for an expert given access. Irreproducible: key details incomplete or key resources not furnished. Do not reject mainly for missing code, data or full proofs. - Convincing - Credible - Irreproducible ### 8. Ethical Concerns Type: free text. "No", or "Yes" with a summary of the ethical issues and how the paper's ethics statement handles them (a Yes triggers senior-PC screening). ### 9. Expertise Type: single choice. How well the paper aligns with your expertise. - Expert - Knowledgeable - Somewhat knowledgeable - Not knowledgeable ### 10. Confidence Type: score, 1–4. How confident you are in your evaluation. - **4**: Very confident. I have checked all points of the paper carefully. I am certain I did not miss any important aspects. - **3**: Confident. I tried to check the key points in the paper. It is unlikely, though conceivable, that I missed some important aspects. - **2**: Somewhat confident. I was unable to carefully check some of the details (proof, novelty, experimental design …) but I am able to defend my evaluation of the paper. It is quite likely that I missed some important aspects. - **1**: Not confident. My evaluation is an educated guess. ## Area chair meta-review IJCAI ACs write the meta-review, a senior AC cross-checks it, and SPC/ACs rate review quality (Exceeding / Met / Failed to meet expectations). Recommendation options: Accept · Reject ## IJCAI reproducibility checklist 1. Theory: assumptions and claims stated formally; theoretical tools cited; proof sketches; full proofs 2. Datasets: novel data in an appendix or public; existing data cited and public; private data described 3. Experiments: code in an appendix or public (or reasons given); hyperparameter search ranges and selection criterion; final hyperparameters; compute infrastructure 4. New algorithms: outline or pseudocode ## LLM policy IJCAI-ECAI 2026 lets reviewers use generative AI only to improve the readability of their own review text, after removing anything identifying or confidential; submissions or their technical content must never be uploaded to tools without confidentiality guarantees. Official policy: https://2026.ijcai.org/principles-of-peer-review-and-recognition/ --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=ijcai (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # KDD 2026 review form > Simulate KDD 2026 Research Track reviews: Paper Summary, Strengths, Weaknesses, Questions for the rebuttal, and 1–4 ratings for Relevance, Novelty, Technical Quality, Presentation, Reproducibility and Reviewer Confidence. - Venue: ACM SIGKDD Conference on Knowledge Discovery and Data Mining (Research Track) - Form status: modeled, not official. KDD does not publish its OpenReview form. Fields and most scale labels are taken verbatim from public KDD 2026 (Cycle 2) and KDD 2025 reviews; the level-1 labels of every scale and the level-2 labels of Relevance and Reviewer Confidence were not seen and are ours. No overall-score field is visible to authors, so Technical Quality is used as the headline score. - Official source: https://kdd2026.kdd.org/research-track-call-for-papers/ - Overall score: Technical Quality, 1–4 - Rough acceptance bar (our simulation): mean ≥ 3 - Web page: https://reviewpaper.ai/venues/kdd ## How reviewers read the paper KDD 2026 runs two cycles (Feb and Aug) on OpenReview with at least three reviews per paper, overseen by ACs and SACs. Reviewers write a summary, numbered strengths and weaknesses, check whether a resubmission addressed earlier reviews, pose rebuttal questions, and rate Relevance, Novelty, Technical Quality, Presentation and Reproducibility on 1–4 (High / Moderate / Low) plus confidence. ACs weigh technical merit, originality, potential impact, execution, presentation, related work, reproducibility and ethics. February submissions can receive a Resubmit decision for the next cycle. - Number strengths and weaknesses; each weakness concrete and actionable. - Ratings must be consistent with the written assessment; say in the rebuttal questions which rating could move. - Reviews of accepted papers become public; keep the tone professional. **Where the bar sits.** KDD 2026 Cycle 1 accepted 256 of 1,215 research-track papers (21%; KDD 2025: 18.7%). KDD has no overall score visible to authors; Novelty and Technical Quality act as the headline ratings. Shared averages of accepted papers cluster around 3 (Moderate) on each 1–4 axis, so a Technical Quality mean ≥ 3 with Novelty ≥ 3 is competitive and ~2.6–3 borderline. ## Reviewer form, field by field ### 1. Paper Summary Type: free text. What the paper does and claims, in your own words. ### 2. Paper Strengths Type: list of points. Numbered strengths: technical merit, originality, potential impact, execution, presentation. ### 3. Paper Weaknesses Type: list of points. Numbered weaknesses, each specific and substantiated: technical flaws, missing baselines or related work, reproducibility gaps, scope of evaluation. ### 4. Resubmission Type: free text. If the paper is a resubmission from the previous KDD cycle, did it address the earlier reviews? Otherwise "N/A". ### 5. Questions And Suggestions For Rebuttal Type: list of points. Questions whose answers could change your ratings; say which rating would move. ### 6. Relevance Type: score, 1–4. Fit to the KDD Research track and breadth of interest. - **4**: High - The work is relevant to the Research track of KDD and is of broad interest to the community - **3**: Moderate - The work is somewhat relevant to the Research track of KDD and is of narrow interest to a sub-community - **2**: Low - The work is only marginally relevant to the Research track of KDD - **1**: Very Low - The work is not relevant to the Research track of KDD ### 7. Novelty Type: score, 1–4. Originality of the ideas relative to prior work. - **4**: High - The paper offers groundbreaking and transformative ideas or approaches that substantially advance the field or open up entirely new areas of research. The level of innovation is high, leading to major advancements and potentially inspiring further research and development. - **3**: Moderate - The paper introduces a new and interesting idea or approach that adds value to the field. The contribution is original and represents an advancement of existing knowledge, demonstrating solid innovation and creativity. - **2**: Low - The ideas are relatively minor and largely incremental. The work builds heavily on existing research. - **1**: Very Low - The work offers no meaningful novelty; its ideas are already known. ### 8. Technical Quality Type: score, 1–4. Rigor of the methodology and analysis; reliability of the results. Used here as the headline score. - **4**: High - The paper exhibits a high level of technical quality with a rigorous and well-executed methodology and analysis. The results are highly reliable, well-supported, and thorough. The work demonstrates technical excellence and sets a high standard for quality in the field. - **3**: Moderate - The paper demonstrates solid technical quality with a sound methodology and thorough analysis. The results are reliable and well-supported. There may be minor issues, but they do not significantly undermine the overall quality. The work is competently executed and meets acceptable standards. - **2**: Low - The paper has several technical weaknesses, such as minor methodological flaws, insufficient analysis, or unsupported conclusions. While the work shows some level of competence, it lacks thoroughness and precision. Improvements are necessary for it to be considered robust. - **1**: Very Low - The paper has major technical flaws that invalidate its main conclusions. ### 9. Presentation Type: score, 1–4. Organization, writing and figures. - **4**: High - The paper is well-organized and very clear. The writing is precise, engaging, and free of errors. Figures and tables are well-designed and seamlessly integrated into the text, enhancing the reader's comprehension. The presentation is polished and professional, making the paper a pleasure to read and understand. - **3**: Moderate - The paper is organized and generally clear. The writing is mostly free of grammatical and typographical errors, making it easy to read. Figures and tables are effectively used to support the text. The presentation facilitates understanding and conveys the key points effectively. - **2**: Low - The paper has noticeable issues with clarity and coherence. The writing may contain several grammatical and typographical errors. Figures and tables are present but may not be well-integrated or effectively used. The presentation allows for understanding but requires effort from the reader. - **1**: Very Low - The paper is poorly organized and hard to understand. ### 10. Reproducibility Type: score, 1–4. Completeness of method/data description, code and data availability. The authors' artifact pledge may be considered positively. - **4**: High - The paper offers a comprehensive and precise description of the methods, data, and procedures. Supplementary materials, including datasets and code, are complete, well-documented, and easily accessible. Reproducing the results would be straightforward and require minimal additional effort, ensuring high reproducibility. - **3**: Moderate - The paper provides a clear and detailed description of the methods, data, and procedures used. Supplementary materials, such as datasets and code, are available and sufficiently documented. Reproducing the results would be feasible with the provided information, though some effort may still be required. - **2**: Low - The paper includes some information about the methods, data, and procedures, but key details are missing. There may be supplementary materials, but they are incomplete or unclear. Reproducing the results would require significant effort and additional information. - **1**: Very Low - The results cannot be reproduced from the information provided. ### 11. Reviewer Confidence Type: score, 1–4. How well you know the area and how carefully you checked the paper. - **4**: High - The reviewer is an expert in the subject area and has extensive knowledge of the research methods and context of the paper. They are highly confident in their ability to provide an accurate and thorough assessment. Their evaluation is based on deep expertise and a comprehensive understanding of the work. - **3**: Moderate - The reviewer has a good understanding of the subject area and is familiar with the research methods and context of the paper. They feel confident in their ability to accurately assess the quality and significance of the work. Their evaluation is based on a solid grasp of the content and context. - **2**: Low - The reviewer has some familiarity with the subject area but may have missed aspects of the methods or related work. - **1**: Very Low - The paper is outside the reviewer's expertise; the evaluation is an educated guess. ### 12. Ethics Review Flag Type: single choice. Does the paper need ethics review? - No - Yes ### 13. Ethics Review Description (optional) Type: free text. If flagged, describe the ethical concern; otherwise "N/A". ## Area chair meta-review KDD ACs write the meta-review, SACs oversee, PC chairs decide. Recommendation options: Accept · Resubmit · Reject ## KDD submission requirements 1. Double-blind: anonymized paper and code links (anonymized GitHub link optional but highly recommended) 2. Artifact pledge (shown to reviewers; reneging after acceptance can lead to retraction) 3. Resubmissions from the previous cycle: previous forum id plus a one-page summary of changes as the first PDF page 4. Use of generative AI disclosed in the submission form ## LLM policy KDD 2026 requires any use of generative AI during reviewing to be disclosed in the review form, and forbids uploading any passage of a paper under review to a generative AI tool. Official policy: https://kdd2026.kdd.org/research-track-call-for-papers/ --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=kdd (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # WWW 2027 review form > Simulate The Web Conference (WWW 2027) research-track reviews: Summary, relevance to the Web, Strengths, Weaknesses, Questions, the −3 to +3 overall evaluation and confidence. - Venue: The ACM Web Conference (WWW) - Form status: modeled, not official. The Web Conference does not publish its review form, and WWW 2027 moves from EasyChair to OpenReview. The overall-evaluation scale (−3 to 3) follows the EasyChair form seen in a public WWW 2026 review (only "2 (accept)" and "1 (weak accept)" were seen; the rest is EasyChair's default). The remaining fields are modeled on the official decision factors. - Official source: https://www2026.thewebconf.org/calls/research-tracks.html - Overall score: Overall Evaluation, -3–3 - Rough acceptance bar (our simulation): mean ≥ 1 - Web page: https://reviewpaper.ai/venues/the-web-conference ## How reviewers read the paper The Web Conference decides from technical merit, originality, potential impact, quality of execution, quality of presentation, related work, reproducibility and ethics, used by ACs and SACs to make recommendations; PC chairs decide. Papers must be relevant to the Web and their track (merely using Web data is out of scope). The rebuttal is written once and is not interactive. WWW 2027 reviews on OpenReview. - Ground every judgment in the official decision factors (merit, originality, impact, execution, presentation, related work, reproducibility, ethics). - Check Web relevance explicitly; do not penalize a paper for the venue fit of its dataset alone. - Professional and specific; questions should be answerable in a one-shot rebuttal. **Where the bar sits.** WWW 2026 accepted about 676 of 3,370 research submissions (~20%). No score statistics are public; on the −3…3 scale a mean of ≥ 1 (weak accept) is a rough proxy for acceptance, 0.25–1 borderline. ## Reviewer form, field by field ### 1. Summary Type: free text. What the paper does and claims, in your own words. ### 2. Relevance To The Web And To The Track Type: free text. Is the paper's relevance to the Web and to its track stated on the first page and convincing? Papers that merely use a Web dataset or API are out of scope. ### 3. Resource Paper Type: single choice. Is the main contribution a resource (dataset, benchmark, tool)? - no - yes ### 4. Strengths Type: list of points. Technical merit, originality, potential impact, quality of execution and presentation, each substantiated. ### 5. Weaknesses Type: list of points. Specific weaknesses in technical merit, originality, execution, presentation, related work, reproducibility or ethics, most important first. ### 6. Questions For The Authors Type: list of points. Questions for the (non-interactive) rebuttal whose answers could change your evaluation. ### 7. Overall Evaluation Type: score, -3–3. Overall recommendation. - **3**: strong accept - **2**: accept - **1**: weak accept - **0**: borderline paper - **-1**: weak reject - **-2**: reject - **-3**: strong reject ### 8. Reviewer's Confidence Type: score, 1–5. How familiar you are with the area and how carefully you checked the paper. - **5**: expert - **4**: high - **3**: medium - **2**: low - **1**: none ## Area chair meta-review WWW ACs and SACs recommend; PC chairs decide. The meta-review form is not public. Recommendation options: Accept · Reject ## WWW submission requirements 1. Relevance to the Web and to the chosen track stated on the first page 2. 8 pages of main content, at most 12 pages including references and appendix 3. Double-blind; self-citations in the third person 4. Ethics/consent section with IRB details where human participants are involved 5. Code and data released after review (encouraged; optional ACM Artifacts Available badge) ## LLM policy The WWW 2026 industry track required reviewers to commit not to rely on LLMs and forbade uploading paper content to any LLM; the research-track call states no reviewer rule, and ACM's peer-review policy forbids uploading confidential submissions to generative AI. Official policy: https://www2026.thewebconf.org/calls/industry.html --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=the-web-conference (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # SIGIR 2026 review form > Simulate SIGIR 2026 full-paper reviews: relevance to the conference, Relevance / Novelty / Technical Soundness / Presentation ratings, Strengths, Weaknesses, an Overall Recommendation (−3 to +3) with justification, and confidence. - Venue: ACM SIGIR Conference on Research and Development in Information Retrieval - Form status: modeled, not official. SIGIR does not publish its review form. Field order and the values seen (e.g. "-1 poor" … "2 excellent", "1 weak accept", "0 borderline") come from public SIGIR 2026 Short Papers reviews on OpenReview; the Full Papers form is assumed to match, and the unseen ends of the overall and confidence scales are ours. - Official source: https://sigir2026.org/en-AU/pages/submissions/full-papers-track - Overall score: Overall Recommendation Score, -3–3 - Rough acceptance bar (our simulation): mean ≥ 1 - Web page: https://reviewpaper.ai/venues/sigir ## How reviewers read the paper SIGIR 2026 moved to OpenReview. Each paper gets at least three PC reviews and a senior PC meta-review; decisions weigh novelty, technical depth, elegance, practical or theoretical impact, and presentation. Reviewers rate relevance, novelty, technical soundness and presentation, list strengths and weaknesses, and justify an overall recommendation. Generic, empty, insulting, late or LLM-generated reviews are "highly irresponsible" and can lead to sanctions. - Fair, thoughtful and respectful; never generic or empty. - Check experimental rigor explicitly: baselines, statistical significance, generalizability, reproducibility. - Back "done before" claims with citations. **Where the bar sits.** SIGIR full papers are historically accepted at about 20–21%; 2026 figures are not published. On the −3…3 scale a mean of ≥ 1 (weak accept) is a rough proxy for acceptance, 0.25–1 borderline. ## Reviewer form, field by field ### 1. Relevance To Conference Type: free text. Is this an information retrieval paper of interest to SIGIR? Explain briefly. ### 2. Relevance Score Type: score, -1–2. Fit to SIGIR. - **2**: excellent - **1**: good - **0**: fair - **-1**: poor ### 3. Novelty Type: score, -1–2. What is new beyond prior work; cite prior work for any "done before" claim. - **2**: excellent - **1**: good - **0**: fair - **-1**: poor ### 4. Technical Soundness Type: score, -1–2. Experimental design, baselines, statistical analysis, claims within the evidence. - **2**: excellent - **1**: good - **0**: fair - **-1**: poor ### 5. Quality Of Presentation Type: score, -1–2. Clarity of motivation, scope and results. - **2**: excellent - **1**: good - **0**: fair - **-1**: poor ### 6. Strengths Type: list of points. Specific strengths, substantiated with evidence from the paper. ### 7. Weaknesses Type: list of points. Specific weaknesses: unsuitable baselines, unreliable or non-generalizable results, evaluation mismatched to the research questions, missing statistical tests, insufficient detail to reproduce. ### 8. Additional Notes (optional) Type: list of points. Minor comments and suggestions. ### 9. Overall Recommendation Type: free text. Justify the overall score: the decisive strengths and weaknesses, weighing novelty, technical depth, elegance, impact and presentation. ### 10. Overall Recommendation Score Type: score, -3–3. Overall recommendation. - **3**: strong accept - **2**: accept - **1**: weak accept - **0**: borderline - **-1**: weak reject - **-2**: reject - **-3**: strong reject ### 11. Nominate For Best Paper Type: single choice. Nominate only exceptional papers. - 0 no - 1 yes ### 12. Reviewer Confidence Type: score, 1–5. How certain you are about your assessment. - **5**: I am absolutely certain that my evaluation is correct. - **4**: I am confident, but not absolutely certain that my evaluation is correct. - **3**: I am fairly confident that my evaluation is correct. - **2**: I am fairly confident that my evaluation is correct, but I have significant doubts. - **1**: My evaluation is an educated guess. ## Area chair meta-review SIGIR senior PC members write the meta-review; PC chairs decide. Recommendation options: Accept · Reject ## SIGIR checklist to strengthen an IR paper 1. Why the problem matters is clearly motivated 2. What is new beyond prior work is explicit 3. Claims stay within the evidence; scope (what was and was not done) is explicit 4. Research decisions are justified 5. Experimental design and scale fit the study; suitable baselines 6. Evaluation matches the research questions; proper statistical analysis 7. Enough detail, data and code to reproduce 8. Results presented clearly and discussed thoughtfully ## LLM policy SIGIR follows the ACM Peer Review Policy and requires that referees not use AI systems to create their reviews; LLM-generated reviews are treated as highly irresponsible. Official policy: https://sigir2026.org/en-AU/pages/submissions/submission-policies-and-information --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=sigir (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # ICASSP 2026 review form > Simulate ICASSP-style reviews: confidence, Importance/Relevance, Novelty, Technical Correctness, Experimental Validation, Clarity and References on 1–4, the four-level Overall Evaluation (Definite reject … Definite accept) and a detailed assessment for the authors. - Venue: IEEE International Conference on Acoustics, Speech and Signal Processing - Form status: modeled, not official. ICASSP does not publish its review form. The criteria and their order come from the official ICASSP 2025 reviewer instructions and the 2026/2027 review criteria; numbered labels are from a public ICASSP 2023 review where seen, and the remaining levels (and the paper-type and confidence options) are ours. - Official source: https://2025.ieeeicassp.org/editorial-procedures/ - Overall score: Overall Evaluation Of This Paper, 1–4 - Rough acceptance bar (our simulation): mean ≥ 2.75 - Web page: https://reviewpaper.ai/venues/icassp ## How reviewers read the paper ICASSP papers get at least three reviews on a fixed list of criteria (importance, novelty, technical correctness, experimental validation, clarity, references), an overall evaluation and a detailed assessment shared with the authors. Extreme ratings carry meaning: "Irrelevant", "Has been done before", "Has major problems", "Insufficient validation" and "Does not cite relevant work" each imply rejection and must be justified. The optional rebuttal goes to area and track chairs, not the reviewers. Review is single-anonymous (authors are visible to reviewers). - Write "the paper", not "you" or "the authors"; no sarcasm. - Scores must match the comments; use the whole scale. - "Done before" or "well known" must come with references. - Ignore minor formatting problems; be specific about writing issues. - Do not ask the authors to cite your own work unless directly relevant. **Where the bar sits.** ICASSP 2026 accepted close to 4,800 of ~11,200 regular submissions (~43%). TPC chairs set the rate; no thresholds are published. On the 1–4 overall scale a mean of ≥ 2.75 (mostly Marginal accept or better) is a rough proxy for acceptance. ## Reviewer form, field by field ### 1. Classify The Type Of Paper Type: single choice. The amount of experimental validation expected depends on the paper type. - Theoretical - Algorithmic / methodological - Experimental / application - Dataset / resource - Survey / overview ### 2. Importance/Relevance Type: score, 1–4. "Of limited interest" only if novelty, clarity and correctness are excellent; "Irrelevant" implies rejection and must be justified thoroughly. - **4**: Of broad interest - **3**: Of sufficient interest - **2**: Of limited interest - **1**: Irrelevant ### 3. Originality/Novelty Type: score, 1–4. "Very original" opens new directions. "Has been done before" implies rejection and requires references. - **4**: Very original - **3**: Moderately original - **2**: Minor originality - **1**: Has been done before ### 4. Technical Correctness / Theoretical Development Type: score, 1–4. List any errors found. "Has major problems" implies rejection. - **4**: Definitely correct - **3**: Probably correct - **2**: Has minor errors - **1**: Has major problems ### 5. Experimental Validation Type: score, 1–4. Judge against the paper type. "Insufficient validation" implies rejection. - **4**: Sufficient validation / theoretical paper - **3**: Limited but convincing - **2**: Lacking in some respects - **1**: Insufficient validation ### 6. Clarity Of Presentation Type: score, 1–4. Be specific about writing problems; ignore minor formatting issues. - **4**: Very clear - **3**: Clear enough - **2**: Difficult to understand in places - **1**: Very difficult to understand ### 7. Reference To Prior Work Type: score, 1–4. "References missing": list them. "Does not cite relevant work" implies rejection; use only when the missing work is well known. - **4**: Excellent references - **3**: References adequate - **2**: References missing - **1**: Does not cite relevant work ### 8. Overall Evaluation Of This Paper Type: score, 1–4. Use the whole scale; marginal scores only when the paper really is in the middle. A 1 or 2 must be justified. - **4**: Definite accept - **3**: Marginal accept - **2**: Marginal reject - **1**: Definite reject ### 9. Award Quality Type: single choice. Should the paper be considered for a best-paper award? - No - Yes ### 10. Strengths Type: list of points. Part of "Detailed Assessment Of The Paper". Specific strengths, consistent with the scores above. ### 11. Weaknesses Type: list of points. Part of "Detailed Assessment Of The Paper". Specific weaknesses and how to fix them; every low score must be explained here. Cite references for any "well known" claim. ### 12. Confidence Type: score, 1–4. How confident you are in your evaluation of this paper. - **4**: Expert in the area - **3**: Knowledgeable - **2**: Some familiarity - **1**: Not familiar with the area ## Area chair meta-review ICASSP meta-reviewers and area chairs recommend to track chairs; TPC chairs give final approval. Recommendation options: Accept · Reject ## ICASSP submission requirements 1. 4 pages of content plus an optional 5th page for references only 2. Generative-AI content disclosed in the acknowledgments (system and sections); grammar/editing use exempt 3. At most 9 submissions per author ## LLM policy IEEE policy forbids processing any content of a manuscript under review through a public AI platform (a confidentiality breach). ICASSP 2026 organizers report using the rebuttal phase to identify AI-generated reviews. Official policy: https://journals.ieeeauthorcenter.ieee.org/become-an-ieee-journal-author/publishing-ethics/guidelines-and-policies/submission-and-peer-review-policies/ --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=icassp (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # Interspeech 2026 review form > Simulate Interspeech-style reviews: confidence, appropriateness, novelty, technical correctness, clarity, reproducibility and references on 1–4, a six-level overall recommendation and detailed comments for the authors. - Venue: Interspeech (Annual Conference of the International Speech Communication Association) - Form status: modeled, not official. Interspeech's CMT review form is not public. The criteria list follows a paper that analysed Interspeech 2019 review data and the official long-paper criteria; a few labels come from a public 2023 Show & Tell review. Most scale labels are ours. - Official source: https://www.isca-archive.org/interspeech_2020/stappen20_interspeech.pdf - Overall score: Overall Recommendation, 1–6 - Rough acceptance bar (our simulation): mean ≥ 3.75 - Web page: https://reviewpaper.ai/venues/interspeech ## How reviewers read the paper Interspeech reviews (double-blind, CMT) rate appropriateness, novelty, technical correctness, clarity, reproducibility and references, give an overall recommendation and write detailed comments (overall impression, novelty, clarity, what to improve) that justify the scores. Long papers are judged on depth of contribution, originality, technical correctness, clarity, quality of references and potential long-term impact. Meta-reviewers consolidate the reviews. - Detailed comments must justify every score (at least ~120 words). - Do not penalize non-native English as long as the content is clear. - Constructive and specific; name missing references. **Where the bar sits.** Interspeech accepts roughly 50% of regular papers (official special-sessions call); the new long-paper track targets below 30%. On this 1–6 scale a mean of ≥ 3.75 (Weak Accept or better) is a rough proxy for acceptance. ## Reviewer form, field by field ### 1. Appropriateness Of Topic For The Conference Type: score, 1–4. Fit to speech science and technology. Explain a 1 or 2 in the comments. - **4**: Very appropriate - **3**: Appropriate - **2**: Marginally appropriate - **1**: Not appropriate ### 2. Novelty And Originality Type: score, 1–4. Originality relative to prior speech/audio work; cite the prior work for low scores. - **4**: Very novel and/or innovative - **3**: Novel - **2**: Somewhat incremental - **1**: Not novel ### 3. Technical Correctness Type: score, 1–4. Soundness of methods, experiments and statistics. - **4**: Technically sound - **3**: Mostly sound, minor issues - **2**: Some significant issues - **1**: Major technical flaws ### 4. Clarity Of Presentation Type: score, 1–4. The English need not be flawless; judge whether the content is understandable. - **4**: Very clear - **3**: Clear enough, could benefit from some revision - **2**: Difficult to follow in places - **1**: Unclear ### 5. Reproducibility Type: score, 1–4. Data, setup, hyperparameters and code availability. - **4**: Easily reproducible - **3**: Mostly reproducible - **2**: Hard to reproduce - **1**: Not reproducible ### 6. Quality Of References Type: score, 1–4. References should be peer-reviewed where possible; name missing work. - **4**: Excellent - **3**: Adequate - **2**: Some important references missing - **1**: Inadequate ### 7. Overall Recommendation Type: score, 1–6. Overall recommendation, consistent with the sub-scores. - **6**: Definite Accept: I feel strongly that this paper must be accepted - **5**: Accept - **4**: Weak Accept - **3**: Weak Reject - **2**: Reject - **1**: Definite Reject ### 8. Strengths Type: list of points. Part of "Detailed Comments For Authors". Overall impression, novelty/originality and innovation, with specifics. ### 9. Weaknesses And Suggested Improvements Type: list of points. Part of "Detailed Comments For Authors". What to improve, with specifics. Together with the strengths, at least 120 words, justifying the scores. ### 10. Reviewer Confidence Type: score, 1–4. Selecting a value confirms you are competent to review the paper. - **4**: Very Confident - I consider myself an expert in this area - **3**: Confident - I am familiar with this area - **2**: Somewhat confident - I have some familiarity with this area - **1**: Not confident - this area is outside my expertise ## Area chair meta-review Interspeech meta-reviewers consolidate reviews for the TPC; the form is not public. Recommendation options: Accept · Reject ## Interspeech author checklist 1. ISCA Code of Ethics, pre-prints policy and conference policy read 2. At most 6 pages (regular) or 10 (long paper); last pages only for acknowledgements, AI disclosure and references 3. Fully anonymized, including PDF metadata, self-citations and supplementary repositories 4. Generative AI used only for editing and polishing, reported in the dedicated section 5. Scientific reporting checklist answered in the submission form ## LLM policy No Interspeech reviewer LLM rule was found. The ISCA Code of Ethics allows authors to use generative AI only for editing and polishing, with disclosure; treat submissions as confidential and do not upload them to LLMs. Official policy: https://interspeech2026.org/en-AU --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=interspeech (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # ACM MM 2026 review form > Simulate ACM Multimedia 2026-style reviews: Paper Summary, multimedia Suitability, Strengths, Weaknesses, the six-level Preliminary Rating (Reject … Accept), confidence and award nomination. - Venue: ACM International Conference on Multimedia - Form status: modeled, not official. Fields and labels are verbatim from the official ACM MM 2024 reviewer guidelines (the last edition that published the form); the 2025–2026 guidelines do not show the form, so they may have changed. The 1–6 numbering of the rating is ours. - Official source: https://2024.acmmm.org/reviewer_guidelines - Overall score: Preliminary Rating, 1–6 - Rough acceptance bar (our simulation): mean ≥ 4 - Web page: https://reviewpaper.ai/venues/acm-multimedia ## How reviewers read the paper ACM MM reviewers judge novelty (what the paper contributes and its value to the community), scientific rigor (experiment design, reproducibility, support for claims, code/data release) and multimedia suitability, separating minor camera-ready fixes from major flaws. Papers on new problems should not be devalued for lacking a new algorithm; arXiv is not a publication. The rebuttal cannot add new experiments. Final reviews and responses become public on OpenReview. - Golden rule: write the review you would like to receive. - Critique "the paper"; do not address the authors as "you". - No vague "it is well known that…" claims; cite specific peer-reviewed work (your own only rarely). - Do not request new experiments for the rebuttal; say clearly what should be clarified. - Neutral proof-reading requests; avoid "native English speaker" wording. **Where the bar sits.** ACM MM 2025 accepted 1,250 of 5,330 research papers (23.5%; 2024: 26.2%). No score thresholds are public; on the 1–6 scale a mean of ≥ 4 (Borderline Accept) is a rough proxy for acceptance. ## Reviewer form, field by field ### 1. Paper Summary Type: free text. About five sentences on what the paper does and claims. ### 2. Suitability Type: single choice. Choose one. ACM MM prefers multimodal/multimedia research. - unsuitable for the MM venue as it involves unimodal/unimedia processing - unimodal/unimedia in nature but of sufficient interest to the multimedia community - definitely suitable for MM as the work demonstrates multimedia processing ### 3. Paper Strengths Type: list of points. Soundness, significance and novelty, and relevance to multimedia; value to the community. ### 4. Weaknesses Type: list of points. Same axes. Separate minor errors fixable for camera-ready from major flaws that warrant rejection. Check scientific rigor: experiment design, reproducibility, support for claims, code/data release. ### 5. Preliminary Rating Type: score, 1–6. Preliminary rating before the rebuttal; a final recommendation follows the discussion. - **6**: Accept - This submission should definitely be presented at ACM Multimedia. - **5**: Weak Accept - I would need strong arguments to reject this submission. - **4**: Borderline Accept - I lean towards acceptance, but I could be convinced otherwise. - **3**: Borderline Reject - I lean towards rejection, but I could be convinced otherwise. - **2**: Weak Reject: I would need strong arguments to accept this submission. - **1**: Reject: This submission should definitely not be presented at ACM Multimedia. ### 6. Confidence Score Type: score, 1–4. How certain you are about your assessment. - **4**: Very Confident: You are absolutely certain about your assessment. You are very familiar with the related work. - **3**: Confident: You are fairly confident in your assessment. It is possible that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. - **2**: Somewhat Confident: You are confident in your assessment, but not absolutely certain. It is not impossible that you did not understand some parts of the submission or that you are unfamiliar with some pieces of related work. - **1**: Not Confident: Your assessment is an educated guess. The submission is not in your area or the submission was difficult to understand. ### 7. Award Type: single choice. Choose one. - No: This paper should not be considered for an outstanding paper award. - Yes: This paper should be considered for an outstanding paper award. ## Area chair meta-review ACM MM ACs summarize each paper's strengths and weaknesses and make a recommendation; the AC form is not public. Recommendation options: Accept (oral) · Accept (poster) · Reject ## ACM MM submission requirements 1. 6–8 pages plus up to 2 pages of references; no appendix in the main file 2. Double-blind; complete OpenReview profiles for all authors 3. Multimodal / multimedia scope (out-of-scope papers may be desk-rejected) 4. At most 10 submissions per author ## LLM policy ACM MM 2026 allows AI tools only to polish review language; clear evidence of AI-generated reviews can get the reviewer's own papers desk-rejected. ACM policy forbids uploading confidential submissions to generative AI. Official policy: https://2026.acmmm.org/site/review-process-guidelines.html --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=acm-multimedia (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # CoRL 2026 review form > Simulate CoRL 2026-style robot-learning reviews: Summary, Strengths, Weaknesses with attention to real-robot evidence, a Limitations check, Questions, a four-level rating and confidence. - Venue: Conference on Robot Learning - Form status: modeled, not official. CoRL's OpenReview form is not public and the 2026 reviewer-instructions page is currently empty. Only the label "weak accept" is confirmed (from the official author instructions); the fields and the four-level rating are modeled on earlier CoRL reviews and secondary summaries of the 2026 instructions. - Official source: https://2026.corl.org/contributions/instruction-for-authors - Overall score: Rating, 1–4 - Rough acceptance bar (our simulation): mean ≥ 2.75 - Web page: https://reviewpaper.ai/venues/corl ## How reviewers read the paper CoRL reviewers write section-by-section reviews with heavy weight on real-robot evidence (simulation-only work is acceptable when the simulator is realistic and the link to real robots is clear) and check that the mandatory Limitations section is honest; a frank limitations section is not a weakness. Papers without any weak-accept-or-better score may be rejected before rebuttal; the one-page rebuttal targets factual errors. - Weigh real-robot evidence heavily, but do not reject realistic simulation work just for lacking hardware. - Do not count an honest limitations section as a weakness. - Ask only questions a one-page factual rebuttal can answer. **Where the bar sits.** CoRL 2026 accepted roughly a third of ~2,100 submissions (secondary source; 2024: ~39%). Papers need at least one weak accept to survive the first round. On this modeled 1–4 scale a mean of ≥ 2.75 is a rough proxy for acceptance. ## Reviewer form, field by field ### 1. Summary Type: free text. The problem, approach and robot-learning contribution, in your own words. ### 2. Strengths Type: list of points. Specific strengths, section by section. Real-robot evidence weighs heavily; simulation-only work is fine when the simulator is realistic and the path to real robots is clear. ### 3. Weaknesses Type: list of points. Specific weaknesses: sim-to-real gap, number and diversity of real trials, baselines, failure analysis, generality of the claims. Each actionable. ### 4. Limitations Type: free text. Is the required Limitations section honest and complete? A frank limitations section is not a weakness. ### 5. Questions Type: list of points. Questions the one-page rebuttal could answer (factual clarifications; no new experiments). ### 6. Rating Type: score, 1–4. Papers without at least one weak accept or above (from a reviewer or the AC) may be rejected in the first round without rebuttal. - **4**: Strong accept - **3**: Weak accept - **2**: Weak reject - **1**: Strong reject ### 7. Confidence Type: score, 1–5. How certain you are about your evaluation. - **5**: The reviewer is absolutely certain that the evaluation is correct and very familiar with the relevant literature - **4**: The reviewer is confident but not absolutely certain that the evaluation is correct - **3**: The reviewer is fairly confident that the evaluation is correct - **2**: The reviewer is willing to defend the evaluation, but it is quite likely that the reviewer did not understand central parts of the paper - **1**: The reviewer's evaluation is an educated guess ## Area chair meta-review CoRL ACs can add reviewers after the rebuttal and recommend decisions; the AC form is not public. Recommendation options: Accept (oral) · Accept (poster) · Reject ## CoRL author requirements 1. 8 pages (9th page at camera-ready), including a mandatory Limitations section 2. Use of generative AI disclosed 3. No citations to non-existent references (checked automatically) 4. Anonymized submission ## LLM policy CoRL 2026 reportedly allows AI help for literature search, clarifying concepts and polishing wording, but forbids uploading manuscripts to tools without enterprise confidentiality and requires reviewers to disclose tools used; judgments must be the reviewer's own. (Official page currently empty; unverified.) Official policy: https://2026.corl.org/contributions/instruction-for-reviewers --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=corl (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference. # MLSys 2026 review form > Simulate MLSys-style reviews: summary, strengths, weaknesses, novelty, technical quality and impact of the ML-systems contribution, questions, an overall merit score and reviewer expertise. - Venue: Conference on Machine Learning and Systems - Form status: modeled, not official. MLSys 2026 reviewed in a private HotCRP form (MLSys 2027 moves to OpenReview) and publishes no reviewer guidelines. The official criteria are novelty, quality, impact and interest; the fields and scales here follow common systems-conference (HotCRP) review forms and are not official. - Official source: https://mlsys.org/Conferences/2026/CallForPapers - Overall score: Overall Merit, 1–5 - Rough acceptance bar (our simulation): mean ≥ 3 - Web page: https://reviewpaper.ai/venues/mlsys ## How reviewers read the paper MLSys research-track papers are judged on a combination of novelty, quality, impact and interest at the intersection of ML and systems; the industry track instead stresses impact, technical quality, clarity and lessons from real-world systems with no novelty requirement. Reviews are double-blind with an author response before decisions. - Judge both the ML and the systems sides: realistic workloads, hardware and baselines matter. - Ask where the gains come from (breakdowns, ablations) rather than only whether they exist. - Specific and constructive; the appendix is optional reading. **Where the bar sits.** MLSys 2026 published about 135 papers; acceptance rates are not published (historically ~20–25%). On this modeled 1–5 scale a mean of ≥ 3 (Weak accept) is a rough proxy for acceptance. ## Reviewer form, field by field ### 1. Paper Summary Type: free text. The system or technique, the workloads it targets, and the claimed gains. ### 2. Strengths Type: list of points. Novelty, quality and impact: new system design or insight, realistic workloads and hardware, strong baselines, meaningful end-to-end gains. ### 3. Weaknesses Type: list of points. Missing or weak baselines, unrealistic setups, unexplained overheads, limited generality, missing ablations or breakdowns of where the gains come from. ### 4. Questions For The Authors Type: list of points. Questions for the author response whose answers could change your evaluation. ### 5. Novelty Type: score, 1–4. Relative to prior systems and ML work. - **4**: Surprisingly new contribution - **3**: New contribution - **2**: Incremental improvement - **1**: Published before ### 6. Overall Merit Type: score, 1–5. Overall recommendation weighing novelty, quality, impact and interest. - **5**: Strong accept - **4**: Accept - **3**: Weak accept - **2**: Weak reject - **1**: Reject ### 7. Reviewer Expertise Type: score, 1–4. Your familiarity with the area. - **4**: Expert - **3**: Knowledgeable - **2**: Some familiarity - **1**: No familiarity ## Area chair meta-review The MLSys meta-review form is not public. Recommendation options: Accept · Reject ## MLSys submission requirements 1. Up to 10 pages two-column plus references; appendix separate (reviewers need not read it) 2. Double-blind (industry track hides author names only) 3. Artifact evaluation is voluntary, after acceptance, and does not affect the decision ## LLM policy No MLSys reviewer LLM policy is published. Authors may not submit papers whose content is generated by LLMs alone (unless illustrative or part of the experiments). Treat submissions as confidential and do not upload them to LLMs. Official policy: https://mlsys.org/Conferences/2026/CallForPapers --- Simulate reviews of your own draft in this venue's form: https://reviewpaper.ai/new?venue=mlsys (3 free reviews). Sample report: https://reviewpaper.ai/sample. reviewpaper.ai is for authors checking their own unpublished drafts. If you are an assigned reviewer, follow your venue's LLM policy: most venues forbid putting a submission you are reviewing into any external AI tool. Not affiliated with any conference.