M6 · Lesson 2 — Writing Research Papers

Writing a Strong
Introduction

Reviewers form their opinion in the first page.
The introduction is where your paper earns the right to be read.

01

M6 · L2 — The Funnel

The standard structure

Every introduction
is a funnel

1. Broad context — why RS matters

2. Specific problem — what existing methods fail at

3. Your approach — what you propose

4. Results preview — headline number

5. Contributions list ← Gold

Each paragraph has a job. Don't let paragraph 1 try to do the work of paragraph 3.

Reviewer reading pattern: First sentence of each paragraph. Abstract. Contributions list. If all three make sense — they continue reading. If any one fails — they start looking for rejection reasons.

02

M6 · L2 — Opening Sentence

The most important sentence in your paper

Your first sentence
sets the entire frame

❌ Generic opening

"Recommender systems have become increasingly important in modern society, helping users navigate the overwhelming amount of information available online."

Every RS paper could open with this. It tells the reviewer nothing about your paper specifically.

✅ Specific opening

"Knowledge graph-augmented recommenders improve warm-user performance by 15–30% but silently degrade on cold-start users — because 95% of items in standard KGs have fewer than 5 triples."

Specific gap, specific data, immediate tension. The reviewer is already curious.

Rule: Your first sentence should describe the problem you solve — not the field you work in. The field is context. The problem is your paper.

03

M6 · L2 — The Contributions List

The most important list in your paper

Writing contributions
that earn their place

Contributions must be:

Specific — not "we propose a novel framework" (every paper does this)
Verifiable — the experiments section must deliver on each bullet
Non-overlapping — don't list the same thing twice in different words
Ordered by importance — the key contribution goes first

Test: After reading your contributions list, does the reviewer know exactly what to look for in the experiments? If not, rewrite until they do.

❌ Padded contributions

"(1) We propose K-RagRec. (2) We design a novel indexing method. (3) We introduce a popularity-selective policy. (4) We conduct comprehensive experiments."

Bullet 4 is not a contribution. Bullets 1–3 are vague.

✅ Specific contributions

"(1) We identify that KG coverage sparsity causes silent cold-start degradation in RAG-based RS. (2) We propose coverage-aware retrieval that adapts to per-item KG density. (3) We prove this achieves a tighter approximation bound than uniform retrieval."

04

M6 · L2 — Reviewer Psychology

Understanding who reads your introduction

What reviewers look for
on page one

First 30 seconds

Is this relevant?

Title, abstract, intro first paragraph. They're deciding whether they're the right reviewer and whether the topic matters.

Next 2 minutes

Is this novel?

Contributions list and related work. They're checking if they already know this, if the gap is real, if the claim is new.

Decision point

Do I believe this?

Methodology + experiments. But their skepticism or acceptance is already shaped by how good your introduction was.

"A weak introduction makes reviewers read the rest looking for problems. A strong introduction makes them read looking for confirmation."

05

M6 · L2 — Key Takeaways

What to remember

01

Funnel structure

Broad → specific → your approach → results → contributions list. Each paragraph has one job.

02

First sentence = the problem

Not "RS is important" — that's the field. Open with the specific gap you're filling.

03

Contributions = contract

Specific, verifiable, non-overlapping, ordered. Experiments must deliver every bullet.

Next: M6 · L3 — Related Work That Positions

06

Writing a StrongIntroduction