M2 · Lesson 1 — Reading Papers Effectively

Anatomy of
an RS Paper

Every RS paper follows the same skeleton.
Knowing what each section really does changes how you read.

01
M2 · L1 — The Structure

Every paper has
the same 6 sections

Abstract
Introduction
Related Work
Methodology
Experiments
Conclusion
02
M2 · L1 — Official vs Reality

What they say vs what they mean

The hidden layer

SectionOfficial PurposeWhat It's Really Doing
AbstractSummarise the paperSell you on reading further
IntroductionMotivate the problemEstablish the gap they claim to fill
Related WorkSurvey prior workArgue why existing work is insufficient
MethodologyPresent the modelThe actual contribution ← slow down here
ExperimentsValidate the modelSelectively show where it works best
ConclusionSummarise findingsRestate + hint at future work
03
M2 · L1 — Abstract

Section 1

The Abstract

Three things always hide in the abstract. Find them in 60 seconds.

01

The Problem

What gap or failure in existing work motivates this paper?

02

The Approach

What did they propose in one sentence?

03

The Claim

What improvement, on what dataset, by how much?

K-RagRec example: Problem = LLM hallucination + noisy RAG · Approach = KG subgraph retrieval with GNN encoder · Claim = 41.6% avg improvement over sub-optimal baseline

04
M2 · L1 — Introduction

Section 2

The Introduction
always follows a funnel

Broad context
"RS is important because..."
Specific problem
"Existing methods fail at X..."
Their solution
"We propose Y which..."
Results preview
Contributions list ← READ THIS
"The contributions list is gold. Authors are forced to be explicit here."

It tells you exactly what they're claiming to have done — and you can hold them accountable to it in the experiments section.

05
M2 · L1 — Related Work

Section 3

Related Work —
skim, don't read

You're only looking for 3 things:

  • Which clusters of prior work do they identify?
  • Which papers do they cite most?
  • What's their argument for why nothing prior solves their problem?
⚠️ WATCH OUT

Authors sometimes misrepresent prior work

To make their contribution look bigger, they may understate what existing methods already do. You'll learn to spot this in L3.

06
M2 · L1 — Methodology

Section 4 — slow down here

Methodology is
the actual paper

Everything else is scaffolding. This is where the contribution lives.

Problem Formulation

How do they formally define the task? What are the inputs and outputs?

Key Equation

Usually one central formula defines the model. Find it and decode it.

Training Objective

What loss function? What optimiser? What's being minimised?

07
M2 · L1 — Experiments

Section 5 — second slow-down zone

Experiments:
read critically

  • Datasets — standard benchmarks or cherry-picked?
  • Baselines — are strong recent methods included?
  • Ablation — what happens when each component is removed?
  • Metrics — do they report the ones where they win?
K-RagRec example

What's missing?

No traditional RS baselines like LightGCN or SASRec. All comparisons are within the KG-RAG-for-LLM family only.

08
M2 · L1 — Reading Between the Lines

Common patterns to watch for

What papers don't
say explicitly

What You SeeWhat It Often Means
"Outperforms all baselines"On these specific datasets, with these hyperparameters
"Simple and efficient"Simpler than the most complex competitor
"Novel framework"We combined two existing ideas
Weak baselines onlyThe strongest competitor wasn't included for a reason
Single dataset experimentsDidn't generalise to other datasets

This isn't cynicism — it's critical reading. Good researchers write this way due to space constraints and competition.

09
M2 · L1 — Key Takeaways

3 things to
remember forever

01

Methodology is the paper

Intro and Related Work are rhetoric. Slow down at Methodology and Experiments.

02

Contributions list = contract

Authors must defend every bullet. Hold them to it in experiments.

03

Read the gaps

What baselines are missing? What datasets weren't tested? Gaps tell you more than results.

Next: M2 · L2 — The 3-Pass Reading Method

10
← → arrow keys to navigate