Writing Paper’s Abstract and Introduction

Vu Nguyen
University of Science, VNU-HCMC

Content adapted from:

• Tao Xie
• Simon Peyton Jones
• Kurt Stirewalt
• Standord’s InfoLab
Outline
◼ Tools
◼ Paper organization
◼ Title
◼ Abstract
◼ Introduction

2/11/2025 2
Tools for writing and managing papers
Purpose Tools
Write paper collaboratively Overleaf, Office 365, Google Docs
Manage and annotate citations and Mendeley, Zotero, Papers, JabRef
papers
Search papers and more Google Scholar, Semantic Scholar, Research Rabbit
Analyze papers, chat with PDF Elicit, ChatGPT, SCISPACE (typeset.io), Research Rabbit,
ChatPDF
Revise papers ChatGPT, Gramarly, Trinka.ai, Gemini
Diagrams, brainstorming Miro, XMind, draw.io
Notes, documents, wiki Notion, Google Docs, Obsidian
Other AI tools askyourpdf.com, ChatPDF, cowriter.ai

2/11/2025 3
Typical Paper Organization
◼ Title, along with names, affiliations, email addresses of authors
◼ Abstract
◼ Introduction
◼ Background
◼ Related work
◼ Proposed approach
◼ Evaluation
❑ Experimental design
❑ Results
❑ Discussions
◼ Threats to validity (optional)
◼ Conclusions
2/11/2025 4
Title – 1
◼ Title is very important for a paper
❑ It is the first thing people notice about your paper
❑ It is the title that people decide to read your paper
◼ It contains keywords, usually at the beginning
◼ If possible, name your method in the title
❑ “SRD: a smart reader of human emotions”

◼ Avoid using uncommon buzzwords in the title

❑ Readers may not understand them
◼ Be specific enough but not too specific about your techniques and
technical terms/jargons
2/11/2025 5
Title – 2
◼ Titles should be concise and informative
❑ “A New Signature File Scheme Based on Multiple-Block Descriptor Files for Indexing Very
Large Data Bases”.
❑ “Signature File Indexes Based on Multiple-Block Descriptor Files”

❑ “An Investigation of the Effectiveness of Extensions to Standard Ranking Techniques for

Large Text Collections”
❑ “Extensions to Ranking Techniques for Large Text Collections”

◼ Titles do not need to be complete sentences

❑ “Duplication of Data Leads to Reduction in Network Traffic”
❑ “Duplication of Data to Reduce Network Traffic”

Source: “How to write a great research paper - Seven simple suggestions” by Simon Peyton Jones

2/11/2025 6
Title – 3
◼ Some reviewers don’t like your claimed approach to be “novel”
❑ “TestEra: A Novel Framework for Automated Testing of Java Programs”
❑ “TestEra: Specification-based Testing of Java Programs Using SAT”

2/11/2025 7
Abstract
◼ Abstract is the second most read section of a paper
❑ The most important part of the paper
❑ It is read by 10 to 500 times more than the paper
◼ It helps readers to decide whether the paper is relevant to them
❑ Readers also decide whether to read further
◼ Opportunity and risk
❑ Good writing entices further reading
❑ Bad writing discourages readers (and search engines) even if your ideas are
good
◼ Readers reading the abstract is more likely interested in your paper
than title readers

2/11/2025 8
Abstract’s Structure - 1
◼ Short common abstract: 50-200 words, 4-10 sentences
◼ Extended abstract: up to 500 words

◼ Abstract is a short summary of a paper, and it typically includes the

following elements
❑ Problem: what is the problem/context you try to address in the paper?
❑ Method: how do you address the problem, your approach to solving the problem
❑ Results: what are the most important results people should know?
❑ Conclusions: what do you want people to take from your solution?

2/11/2025 9
Abstract’s Structure - 2
◼ Some conferences and many journals require a “structured abstract”
❑ For example, Journal of Software and Information Technology uses the following
headings, each starting a separated paragraph
◼ Context, Objective, Method, Results, Conclusions

2/11/2025 10
Abstract’s Structure - 3
◼ Another example
When projects lack sufficient local data to make predictions, they try to transfer information
from other projects. How can we best support this process? In the field of software
engineering, transfer learning has been shown to be effective for defect prediction. This
paper checks whether it is possible to build transfer learners for software effort estimation.
We use data on 154 projects from 2 sources to investigate transfer learning between
different time intervals and 195 projects from 51 sources to provide evidence on the value
of transfer learning for traditional cross-company learning problems. We find that the same
transfer learning method can be useful for transfer effort estimation results for the cross-
company learning problem and the cross-time learning problem. It is misguided to think
that: (1) Old data of an organization is irrelevant to current context or (2) data of another
organization cannot be used for local solutions. Transfer learning is a promising research
direction that transfers relevant cross data between time intervals and domains.

Kocaguneli, E., Menzies, T., & Mendes, E. (2015). Transfer learning in effort estimation. Empirical Software Engineering, 20(3), 813-843.

2/11/2025 11
Abstract’s Structure - 4
◼ Another example
Testing in Continuous Integration (CI) involves test case prioritization, selection, and
execution at each cycle. Selecting the most promising test cases to detect bugs is hard if
there are uncertainties on the impact of committed code changes or, if traceability links
between code and tests are not available. This paper introduces RETECTS, a new method
for automatically learning test case selection and prioritization in CI with the goal to minimize
the round-trip time between code commits and developer feedback on failed test cases. The
RETECTS method uses reinforcement learning to select and prioritize test cases according to
their duration, previous last execution and failure history. In a constantly changing
environment, where new test cases are created and obsolete test cases are deleted, the
RETECTS method learns to prioritize error-prone test cases higher under guidance of a
reward function and by observing previous CI cycles. By applying RETECTS on data
extracted from three industrial case studies, we show for the first time that reinforcement
learning enables fruitful automatic adaptive test case selection and prioritization in CI and
regression testing.
Spieker, H., Gotlieb, A., Marijan, D., & Mossige, M. (2017, July). Reinforcement learning for automatic test case prioritization and selection in continuous integration.
In Proceedings of the 26th ACM SIGSOFT International Symposium on Software Testing and Analysis (pp. 12-22).
2/11/2025 12
Abstract writing advice - 2
◼ Don’t put unexplained or undefined terms whose meanings are not well
known
❑ Solutions: explain them; rephrase them using plain words; not get into too much
detail (without mentioning them).

◼ Minor details or a description of the structure of the paper are usually

inappropriate
◼ Acronyms, mathematical equations, abbreviations, or citations
◼ Rare: if citations are needed, provide full citations.
❑ Ex. Don’t put [1] in the abstract

2/11/2025 13
Abstract writing advice - 3
◼ Many recommend writing the abstract last as it provides a short and
concise summary of the paper
◼ The key message, terms, and numbers have to be consistent across
the abstract, introduction, and conclusion
◼ Your paper may have more than two results and conclusions, only
state the most important one or two in the abstract
◼ Abstract must be stand-alone
❑ Do not cite papers or sections of the paper in the abstract

2/11/2025 14
Abstract writing advice - 4
◼ Abstract must be concise; do not leave readers to guess what you
mean
◼ Avoid any grammatical errors in abstract

Source: https://www.elsevier.com/connect/writing-a-science-paper-some-dos-and-donts

2/11/2025 15
Abstract Writing Exercise
◼ Find and read one paper
❑ Copy the introduction section of the paper to a Google Docs
◼ Work in groups of two each to write an abstract for the paper you read,
in 200 words
◼ For those groups having chosen a topic, you can write an abstract for
your topic
❑ Time: 30 minutes, let’s come back at 4:05
◼ Compare your abstract with the original one

◼ Share your Google Docs with the class

◼ We together review and revise each abstract

2/11/2025 16
Organizing (conference paper)
◼ Title (1000 readers)
◼ Abstract (4 sentences, 100 readers)
◼ Introduction (1 page, 100 readers)
◼ The problem (1 page, 10 readers)
◼ My idea (2 pages, 10 readers)
◼ The details (5 pages, 3 readers)
◼ Related work (1-2 pages, 10 readers)
◼ Conclusions and further work (0.5 pages)

Source: “How to write a great research paper - Seven simple suggestions” by, Simon Peyton Jones, Microsoft Research, Cambridge

2/11/2025 17
Introduction - 1
◼ An introduction can be regarded as an expanded version of the
abstract
❑ Should not just simply restate the abstract
❑ Should not contain the same sentences from the abstract
◼ It describes the paper’s topic, the problem being studied, references to
key papers, the approach to the solution, the scope and limitations of
the solution, and the outcomes.
◼ There needs to be enough detail to allow readers to decide whether or
not they need to read further.

2/11/2025 18
Introduction - 2
◼ Be modest, do not overclaim
❑ But it is good to polish and put your work in a bigger picture
◼ The paper needs to justify any claims made in the introduction
◼ Do not over-criticize other’s work
◼ If you want to claim some unjustified points, it is better to put them in
conclusion or discussion section
◼ Even if so, be careful on wording
❑ “Our approach provides a foundation for this new field.”
❑ “We believe our approach can provide a foundation…”
❑ “We believe our approach has a good potential for providing a foundation …”

2/11/2025 19
Introduction’s Structure - 1
◼ Stirewalt's 5-paragraph rule
1) Introductory paragraph
◼ Very briefly: What is the problem and why is it relevant to the audience attending
*THIS CONFERENCE*?
◼ Moreover, why is the problem hard, and what is your solution? You must be brief
here. This forces you to boil down your contribution to its bare essence and
communicate it directly.

2) Background paragraph
◼ Elaborate on why the problem is hard, critically examining prior work, trying to
tease out one or two central shortcomings that your solution overcomes.

Source: http://www.cse.msu.edu/%7Echengb/Writing/intro-guidelines-stirewalt.txt

2/11/2025 20
Introduction’s Structure - 2
◼ Stirewalt's 5-paragraph rule (cont’d)
3) Transition paragraph
◼ What keen insight did you apply to overcome the shortcomings of other
approaches?
◼ Structure this paragraph like a syllogism: Whereas P and P => Q, infer Q.
4) Details paragraph
◼ What technical challenges did you have to overcome and what kinds of validation
did you perform?
5) Assessment paragraph
◼ Assess your results and briefly state the broadly interesting conclusions that
these results support. This may only take a couple of sentences. I usually then
follow these sentences by an optional overview of the structure of the paper with
interleaved section callouts.

Source: http://www.cse.msu.edu/%7Echengb/Writing/intro-guidelines-stirewalt.txt

2/11/2025 21
Introduction’s Structure - 3
The Stanford InfoLab's patented five-point structure:
1. What is the problem?
2. Why is it interesting and important?
3. Why is it hard? (e.g., why do naive approaches fail?)
4. Why hasn't it been solved before? (or, what's wrong with previous
proposed solutions? How does mine differ?)
5. What are the key components of my approach and results? Also
include any specific limitations.

Source: https://cs.stanford.edu/people/widom/paper-writing.html

2/11/2025 22
Jones’ Structure
◼ One page
❑ Describe the problem
◼ May give an example Use an
❑ State your contributions example
❑ … and that is all to
introduce
the
problem

Source: “How to write a great research paper - Seven simple suggestions” by

Simon Peyton Jones, Microsoft Research, Cambridge
http://research.microsoft.com/en-us/um/people/simonpj/papers/giving-a-talk/giving-a-talk.htm

2/11/2025 23
Molehills not Mountains

Example: “Computer programs often have bugs. It is very important to eliminate

these bugs [1,2]. Many researchers have tried [3,4,5,6]. It really is very important.”

Yawn!
Example: “Consider this program, which has an interesting bug. <brief
description>. We will show an automatic technique for identifying and removing
such bugs”

Cool!

Source: “How to write a great research paper - Seven simple suggestions” by

Simon Peyton Jones, Microsoft Research, Cambridge
http://research.microsoft.com/en-us/um/people/simonpj/papers/giving-a-talk/giving-a-talk.htm

2/11/2025 24
State Your Contributions - 1

◼ Write the list of contributions first

◼ The list of contributions drives the entire paper: the paper
substantiates the claims you have made
◼ Reader thinks “gosh, if they can really deliver this, that’s be
exciting; I’d better read on”

Source: “How to write a great research paper - Seven simple suggestions” by

Simon Peyton Jones, Microsoft Research, Cambridge
http://research.microsoft.com/en-us/um/people/simonpj/papers/giving-a-talk/giving-a-talk.htm

2/11/2025 25
State Your Contributions - 2
◼ You can state like the
following
❑ “The paper makes the following
main contributions: + bulleted
items”
❑ “The study offers several main
contributions: + bulleted items”

❑ It is good for readers to spot out

main contributions

2/11/2025 26
Contributions should be refutable

No! Yes!
We describe the WizWoz system. It is We give the syntax and semantics of a
really cool. language that supports concurrent processes
(Section 3). Its innovative features are...

We prove that the type system is sound, and

We study its properties that type checking is decidable (Section 4)

We have built a GUI toolkit in WizWoz, and

used it to implement a text editor (Section 5).
We have used WizWoz in practice The result is half the length of the Java
version.
Source: “How to write a great research paper - Seven simple suggestions” by
Simon Peyton Jones, Microsoft Research, Cambridge
http://research.microsoft.com/en-us/um/people/simonpj/papers/giving-a-talk/giving-a-talk.htm

2/11/2025 27
Evidence

◼ Your introduction makes claims

◼ The body of the paper provides evidence to support each claim
◼ Check each claim in the introduction, identify the evidence, and
forward-reference it from the claim
◼ “Evidence” can be: analysis and comparison, theorems,
measurements, case studies

Source: “How to write a great research paper - Seven simple suggestions” by

Simon Peyton Jones, Microsoft Research, Cambridge
http://research.microsoft.com/en-us/um/people/simonpj/papers/giving-a-talk/giving-a-talk.htm

2/11/2025 28
Exercise 1

◼ Read one paper published in good conferences

◼ Identify each section name and its purpose

◼ Suggest an outline for your paper

2/11/2025 29
Exercise 2 - Analyze the introduction section

◼ Work on the same paper chosen in the exercise above

◼ Identify sentences concerning the following in the intro section
❑ Context, background
❑ Problem
❑ Why the problem is hard and important
❑ Previous work addressing the problem
❑ Solution/method
❑ Evaluation
❑ Results
❑ Claims
❑ Contributions

◼ Is there anything missing from the intro section that prevents

readers from understanding it?

2/11/2025 30