M6 · Lesson 6 — Writing Research Papers

Responding
to Reviews

A rejection is not the end of a paper.
A rebuttal is a negotiation — and most researchers do it badly.

01
M6 · L6 — Reading Reviews

Before you write a single word

Read the reviews
before reacting to them

Most researchers react emotionally to reviews. That's human. But wait 24 hours, then read again asking:

  • What is the reviewer actually concerned about — not what they wrote literally?
  • Is this concern valid — even if poorly worded?
  • Is this concern about the paper or the research?
  • What is the minimum change that addresses this concern?
What reviewers really mean

"The baselines are weak"

Often means: "I don't believe the improvement is real because I know a stronger method you didn't compare to." Add the method, or explain why it's not comparable.

What reviewers really mean

"The novelty is limited"

Often means: "I can't see why this combination needed a paper." Reframe your contribution — the research may be fine but the framing is failing.

02
M6 · L6 — Rebuttal Structure

The three-part response

Acknowledge → Address
→ Evidence

Part 1 — Acknowledge

Show you understood the concern

"The reviewer raises a valid concern about our evaluation protocol. We acknowledge that..."

Do not start with "We disagree." Even if you disagree — acknowledge first.

Part 2 — Address

Provide the specific response

"To address this, we added experiments with [method] as a baseline. Results show..."

Part 3 — Evidence

Point to concrete support

"See Table X (revised) and Appendix Y. The improvement holds across all three datasets."

❌ Defensive rebuttal

"The reviewer misunderstood our method. We clearly explained in Section 3 that our approach is different from what the reviewer describes."

Telling a reviewer they misunderstood never works.

✅ Constructive rebuttal

"We thank the reviewer for highlighting this. We have clarified Section 3 to make this distinction explicit, and added Figure X which visualises the difference."

03
M6 · L6 — Prioritising Concerns

Not all concerns are equal

What to fix vs
what to argue

Fix these without arguing:

  • Missing baselines the reviewer names specifically
  • Unclear writing or presentation issues
  • Missing ablation studies on key components
  • Incorrect or missing citations

These you can argue — carefully:

  • Reviewer requests a baseline that is architecturally incompatible
  • Reviewer misunderstands your task definition
  • Reviewer compares to a concurrent unpublished paper
"Never argue about whether your improvement is significant. Add the experiment. Let the numbers argue for you."

Prioritisation rule: Fix everything that can be fixed in the revision period. Argue only what genuinely cannot be addressed with more experiments or clearer writing.

04
M6 · L6 — After Rejection

The long game

Rejection is
a data point, not a verdict

Every rejection contains information. Extract it:

  • Which concern appeared in multiple reviews? That's the real problem.
  • Which concern appeared in only one review? That might be noise.
  • Did reviewers understand your contribution? If not — framing problem, not research problem.
  • Did reviewers ask for experiments you can do? Then do them before resubmitting.
The honest question

Is this a framing rejection or a research rejection?

Framing rejection: reviewers couldn't see the contribution. Fix the introduction and related work. Resubmit to a better-fit venue.

Research rejection: the method has a genuine flaw. Fix the method. Don't resubmit the same paper.

Most rejections are framing rejections. The research is fine. The paper didn't help reviewers see it clearly.

05
M6 Complete — Writing Research Papers

Module 6 complete

L1–L2

Contribution + Introduction

Gap → Bridge. 3-sentence formula. Funnel structure. Specific first sentence. Contributions list as contract.

L3–L4

Related Work + Method

Position don't list. Theme over chronology. Problem formulation → Overview → Components → Training.

L5–L6

Novelty + Reviews

Replace "novel" with specifics. Impact = enables + opens. Reviews: Acknowledge → Address → Evidence.

Final thought

The skills in M1–M6 compound. Notation fluency (M1) enables equation writing (M5). Critical reading (M2–M3) builds positioning instincts (M6). Reproduction experience (M4) gives you the credibility to claim novelty (M6-L5). You now have the full toolkit. Use it. 🎯

Next: M7 · Integration — end-to-end analysis of a real RS paper

06
← → arrow keys to navigate
1 / 6