How to write research papers? Ping HU, Kuniaki Saito A papers - - PowerPoint PPT Presentation

How to write research papers?

Ping HU, Kuniaki Saito

A paper’s impact on your career

Our image of the research community

• Scholars, plenty of time on their hands, pouring over your manuscript.

The reality: more like a large, crowded marketplace

Content

★ Before Writing ★ How to Write ★ After Writing

➔ Choose a topic

Before writing a paper

How to choose topics?

• Your own interests.
• Advices from supervisor.
• Follow the community's hot topics.

➔ Literature review

Before writing a paper

How do you reading papers when you write papers?

• Abstract
• Introduction
• Related work
• Experiments
• Methodology
• Conclusion
• Future works
• References

➔ Experiments & Writing

Before writing a paper

Manage your work progress

• Experiments first?
• Writing first?
• At the same time?

➔ Write for submission

Before writing a paper

• Choose a journal/conference/workshop to

submit your paper.

○ Area & Preference ○ Reputation & Accepting rate ○ Timeline ○ Page limit ○ Format requirements ○ Page Charges ...

Content

★ Before Writing ★ How to Write ★ After Writing

Structure of a paper

• Title, Abstract
• Introduction
• Related Work
• Method
• Experiment
• Discussion, Analysis
• Summary, Future Work

1, Which parts should we spend time the most? 2, Which parts should we start to write at first?

Structure of a paper

• Title, Abstract
• Introduction
• Related Work
• Method
• Experiment
• Discussion, Analysis
• Summary, Future Work

Spend the most time, start with here! Depends on paper, journal, conference...

Review: Aspect of a good paper

• Novelty
• Presentation
• Originality
• Application
• Educational (Based on journal or conference)

What should we keep in mind in writing a paper?

• We propose something.
• Others did not do it.
• We did something.
• We did experiments, ours was better.

?

What should we keep in mind in writing a paper?

• A. Tell a story

○ Here is a problem… ○ It is an interesting, unsolved problem ○ Here is my idea (Novelty, Originality) ○ How my idea works (Detail, Experiment) ○ How my idea is different from others

Reader or Reviewer

Introduction

• Summarize a story you made

○ Describe a problem ○ Clearly state your contributions

Method

• Explain a method as if you were speaking to someone

○ Give examples to describe

• Convey intuition → Detail

○ Readers can follow the detail given general intuition ○ Readers can have some understanding even if they skip detail

Revise!

http://www.newcollege.utoronto.ca/student-blog/how-to-write-a-philosophy-paper-10

• practical-tips/

Style of writing and revising

• Roughly writing all sections → Revise each section ⇄ Look at whole parts
• (Introduction ⇄ Revise) → (Method ⇄ Revise) → … (Conclusion ⇄ Revise)
• Write a conclusion? → Next section?

What kind of writing style do you think effective?

Reviewing papers with each other

• A good paper for authors may not be a good one for others
• Advice given by colleagues is effective

○ with the same interest ○ with different interest

Content

★ Before Writing ★ How to write ★ After Writing

➔ Response to peer review

After writing the paper There always are several rounds of reviewing processes before the paper get accepted.

• Conferences:

submission -> peer review->rebuttal -> Final Decision

• Journals:

Submission -> 1st round review -> author response -> 2nd round review -> author response -> …

• > Final Decision

➔ Response to peer review

After writing the paper

You: Here is a faster horse R1: You should have used my donkey R2: This is not a horse, it’s a mule R3: I want a unicorn!

➔ Response to peer review

After writing the paper

You: Here is a faster horse R1: You should have used my donkey R2: This is not a horse, it’s a mule R3: I want a unicorn!

• 1. Take a break to calm down
• Read and understand the reviews
• 2. Give point-to-point reponses.
• Number the comments sequentially
• 3. Provide well-reasoned arguments
• Less emotional statements, but more rational

arguments

• 4. Watch your tone
• 5. Thank the reviewers for their time and effort

Reviewer 1 Comment 1. .... Response: … Comment 2. .... Response: … Reviewer 2 ....

➔ Prepare for publication

After writing the paper

After the paper get accepted

• Submission deadline
• Publisher's requirements.
• Revision and proofreading
• Acknowledgement
• Appendix and source codes
• Pay the page charges

Paper Gestalt

• Jiabin Huang et al. "Deep Paper Gestalt", Arxiv, 2018
• Main Point: Design algorithm to analysis the

pretty-looking layout of paper and the right mix of equations, tables and figures

• Learn a classifier to predict whether a paper should be

accepted or rejected based solely on the visual appearance of the paper

• Safely reject 50% of the bad papers while wrongly

reject only 0.4% of the good papers

Bad papers Confidence Map Good papers

Failing to fill the paper into full pages is a discriminative visual cue for bad paper The top-right corner of the first page is important for understanding the idea.

Bad papers Good papers

Tools

• Google scholar h-index
• DBLP
• Software: Latex/ Overleaf/ Grammarly
• h index = significance?
• # of citation = significance?
• Several influential papers have been rejected once or twice
• Some best papers make little impact
• Never give up in the process

Lessons

Reference

How to write a great research paper

https://www.cis.upenn.edu/~sweirich/icfp-plmw15/slides/peyton-jones.pdf

Thank you!