How to Effectively Address Reviewer Concerns About Insufficient Strong Baseline Comparisons
The article outlines why reviewers question weak baseline choices, shows common ineffective replies, and provides a structured, evidence‑based rebuttal template that acknowledges the issue, details existing baseline coverage, adds missing experiments when possible, and explains any unavoidable limitations.
Problem
Reviewers often state that a manuscript lacks comparison with the latest or strongest baselines, which weakens the experimental evidence.
Why simple “resource‑limited” replies fail
Answering only that time or compute resources prevented inclusion of the baseline shifts focus to resource constraints and does not address the reviewer’s core concerns.
Reviewers care about:
Whether the chosen baselines are representative.
Whether major method categories are covered.
Whether the most relevant or strongest methods are omitted.
Whether conclusions still hold with stronger baselines.
Whether a reasonable explanation is provided if a baseline cannot be added.
Recommended rebuttal structure
First acknowledge the importance of strong baseline comparison, then describe the coverage of the baselines already used, and finally add any feasible experiments.
Thank you for the valuable suggestion. We agree that comparing with stronger baselines would further strengthen the experimental evidence. In the current version we have compared methods B, C, and D, which represent traditional approaches, classic deep‑learning models, and the latest large‑model techniques, respectively. We have now added a comparison with method A as you suggested; the results (see Table X) show that our approach still outperforms A on the XXX metric, confirming that our conclusions hold under a stronger baseline. We will include these results in the revised manuscript and elaborate on the baseline selection rationale.
Core answer checklist
1. What does the current baseline set cover?
Explain that the selected baselines span major categories such as:
Traditional methods.
Classic deep‑learning models.
Latest large‑model approaches.
The most related prior work.
Strong competitors on the target task.
2. Can the reviewer‑suggested baseline be added?
If time permits, directly add the missing experiment and report the key numbers or trends, e.g., “We have added a comparison with method A; the results show a Y% improvement on metric Z.”
We have added a comparison with method A; the results show that our method still achieves superior performance on the XXX metric.
3. If the baseline cannot be added, why?
Provide a concrete explanation rather than a generic “resource‑limited” claim, such as:
Method A requires additional annotated data unavailable for this study, depends on a closed‑source model that cannot be reproduced, or targets a different task setting.
Also reaffirm that the existing baselines (B, C, D) already cover the main technical routes, preserving their representativeness.
One‑sentence takeaway
Do not rely on “insufficient resources” as the main excuse; instead clarify the baseline coverage, add feasible experiments, or give a specific, justified reason why the current baselines remain representative.
GitHub repository for the full tip collection: https://github.com/MLNLP-World/Paper-Rebuttal-Tips
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
Machine Learning Algorithms & Natural Language Processing
Focused on frontier AI technologies, empowering AI researchers' progress.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
