How Not to Get Your Paper Rejected Mainak Chaudhuri Indian Institute of Technology, Kanpur Prolog: Why Publishing is important Must tell your ideas to others Improves the chance of converging to a better idea (hundreds of brains working vs.
"How Not to Get Your Paper Rejected Mainak" is the property of its rightful owner. Permission is granted to
download and print the materials on this website for personal, non-commercial use only, and to display it
on your personal computer provided you do not modify the materials and that you retain all copyright
notices contained in the materials. By downloading content from our website, you accept the terms of this
agreement.
Presentation Transcript
01
How Not to Get Your Paper Rejected Mainak Chaudhuri
Indian Institute of Technology, Kanpur<br>
02
Prolog: Why Publishing is important
Must tell your ideas to others
Improves the chance of converging to a better idea (hundreds of brains working vs. tens)
Helps the community to push the frontline in the right direction
Inspire others to think smart by showing them how to think smart
Let the coming generation(s) remember you for what you accomplished
Caveat: publishing is often overhyped (this is where the counting game takes over)<br>
03
Searching for Ideas Selling a good idea is far easier than selling a not so good idea
Truth be told, there aren’t too many good ideas out there
Read a lot of papers to stay abreast with what people around the globe are thinking
Identify the top conferences and journals in your area and browse the proceedings and periodicals in regular intervals
Reserve time for this from your weekly/monthly plan
Summary: reading good papers holds the key to writing good papers<br>
04
Idea-Execution Cycle The cycle of idea-(execute-improve)+ continues until the results are satisfactory or further gains are marginal
Understand what is satisfactory by the publishing standards
Depending on the standard of results, decide where to publish
Top tier conferences and journals usually demand high standard ideas and results
These are usually medium to highly influential ideas
Be realistic and decide how good your idea is
Sending a not-so-good result to a top conference/journal is a non-starter<br>
05
Draw up an Outline Think about a good story to tell
Draw up the sections and subsections
At every step put yourself in the position of a reader and ask if the story plot is interesting
Put enough time to write a good abstract and introduction
First impression is the final impression
Motivate the problem and summarize your solutions in introduction
If your contribution is building on top of several other prior work, include a related work subsection at the end of introduction; otherwise move the related work section to the end of the paper<br>
06
Give Credit to Others Not acknowledging closely related studies is the biggest sin in research
Know the related work well and review them generously
Importantly, point out how your idea improves previous state-of-the-art
In the evaluation section, compare your proposal quantitatively with a few best known prior arts
Implement them in your infra-structure and evaluate impartially<br>
07
Mid-game (1/3) The middle of the paper goes in explaining your contribution
Crux of the matter
Polish the writing well and make sure that you highlight the major contributions
Usually, this part is easy to write because this is your work
Make sure to show how you got to your end-design step by step
Motivate every addition to your design and tell a good story<br>
08
Mid-game (2/3) Once you have talked about the central ideas, it is time for quantitative evaluation
Furnish enough information about your simulation/evaluation infra-structure so that others can reproduce the results
Very important for the contribution to be useful
Provide exact values of the important parameters that you have used
If you are using some non-conventional/new methodology for evaluating your idea, justify why it makes sense
Use publicly available benchmark suites and furnish enough details about them<br>
09
Mid-game (3/3) Once the infra-structure is conveyed, proceed to present your results
Explain the results carefully highlighting your contributions
Spend time to explain any outliers
This opens up further trains of research
Compare your proposal with a few closely related studies
Explain why you are better
Explain any outliers with more care
Overall, the evaluation should make a convincing story and tell readers that you have spent time to carefully design your experiments<br>
10
Read-Think-Update Read the complete manuscript several times (not back-to-back)
Reserve at least seven days for final polishing
Challenge yourself with hard questions that a reader may ask; see if you have answers to them in the paper
While reading imagine that you know nothing about the work; see if you can follow the train of thoughts described in the paper
Come up with a list of negatives about your paper which may lead to rejection
Address them carefully
Ask colleagues to read and give feedback<br>
11
Typical Outline Abstract
Introduction, Motivation, Related Work
Background
Your Design Proposal
Evaluation Methodology
Results and Analysis
Summary
References
Appendices (if any)<br>
12
Epilog: Rejection Can Help The world is highly competitive
Too many smart people out there
Probability of getting rejected is fairly high (top tier CS conferences have one in ten to one in five acceptance rates)
If your paper is rejected, take the reviews seriously and try to address the concerns
Most reviews are usually good if you tried a top tier conference/journal
Some reviews may be unreasonably critical; don’t get dejected
Revise well and try again<br>
13
How Not to Get Your Paper Rejected Thank you
Mainak Chaudhuri
Indian Institute of Technology, Kanpur<br>
