Friday, August 9, 2013

How to Write an Excellent Technical Conference Paper

One of my roles as an ECE PhD candidate is to write and publish IEEE conference papers. If you've been accepted to an engineering PhD program, your advisor will expect you to publish conference papers on your research. Now, I've seen great conference papers that tell good stories, and I've seen papers that I've tossed aside because they have little merit because I can hardly make head or tails of them. Today, I will give you tips on how to write conference papers that will be excellent because they tell your research story clearly and succinctly.

To begin, you should understand that many technical conferences exist, and some conferences need papers so badly that the technical chairs will accept almost any paper. In other words, you need to choose your conference carefully. Even the Institute of Electrical and Electronic Engineers (IEEE) has poor quality conferences even though the IEEE is regarded for publishing excellent technical content. As you begin writing your paper, ask your PhD advisor to what conference you should submit your paper. He/she will have ideas on which conferences demand excellent papers. For example, in my research field of antenna arrays optimized with genetic and other stochastic algorithms, excellent conferences include

  • IEEE Antennas and Propagation Symposium
  • IEEE Microwave Theories & Techniques Society International Microwave Symposium
  • IEEE Congress on Evolutionary Computation
  • ACM Genetic and Evolutionary Computation Conference
  • IEEE-APS Topical Conference on Antennas and Propagation in Wireless Communications
This list is certainly not complete, nor may it be relevant to your research areas. However, each of these conferences expects high quality papers, and paper submissions are peer reviewed. If your paper is accepted, that means your paper is good.

Second, you need to understand how a conference paper is organized. A conference paper is not the same thing as writing an essay for an english class or a final project report for a capstone course. Conference papers are divided into clear sections as shown in the figures below. 

Diagram showing first page of a sample conference paper 

Diagram showing a final page of a sample conference paper


Conference papers are typically double columned and single spaced with page limits ranging from two pages to eight pages. Each conference has their own paper formatting rules, and you can find those rules on the conference's website. In general, conference papers have several ordered sections following the title and authors lines:
  1. Abstract
  2. Introduction / Background
  3. Methods (Alternatively: Experimental setup, Theory)
  4. Results (Alternatively: Experimental results, Simulation Results)
  5. Conclusion
  6. Bibliography (i.e., references list)

Tip: Download the appropriate paper template from the conference website

Tip: Use LaTex instead of Microsoft Word to write your paper


As you notice, a conference paper will typically have several diagrams and/or figures. I've found Microsoft Word to be tricky when formatting papers because my figures would not stay in one place when I added text to my document, and it's very easy for Word to mess up font formatting even when I used the conference's Word template. Instead, I've found paper formatting to be much easier by using LaTex (pronounced "Lay Tech").  LaTex is a type-setting program commonly used in academia. If you do not already know how to use LaTex, I've found a thorough Latex Tutorial. A version of Latex called MacTex is available for the Macintosh, and MikTex is available for Windows PCs. Once you've installed LaTex on your computer, I suggest that you also install and download TexMaker which is a cross platform LaTex editor that works on top of MacTex / MikTex. I've found TexMaker easier to use that MacTex and MikTex by themselves.

Regarding the paper outline, the abstract summarizes the paper in 100 to 150 words. Second, the introduction / background tells why your research is important, what other researchers have done in that research area, and what you have done previously (if this is not your first paper on the subject). Remember to cite any papers or books you reference in your introduction. You might also include a diagram explaining your research at an upper level. This may be a picture of your project. Third, you describe your test setup or simulations methodology in your methods section. You can name this section differently as noted above. This section typically includes a picture or diagram explaining your research's methodology. It might be a block diagram, an algorithm flowchart, etc. The important thing to note is that this section contains the theoretical meat of your paper. 

Fourth, you discuss your test and/or simulation results in the results section. You must show clear diagrams of your results. Interpret your results and explain why and how they matter. In general, you want to show results that support your research hypotheses. It is acceptable to publish results that disprove your hypotheses, but it's more interesting to prove your own points rather to disprove yourself. Fifth, the conclusion summarizes your paper. Your conclusion can be similar to your abstract, but some authors include future work in their conclusions. Finally, the bibliography lists all of the literature your referenced throughout your paper.

In closing, you must tell a story when you write a conference paper. Have your PhD advisor review your paper before you submit it, as your advisor will know the level of quality required for conference acceptance. In most cases, the conference reviewers will also give you feedback on your paper after you submit it but before they accept it. Do not let their comments get you down, as the reviewers do not know your research as well as you do. Anyway, your paper will be much better after you integrate their comments and suggestions into your revisions.


Best,

Jonathan Becker
ECE PhD Candidate
Carnegie Mellon University
d7662bcda31b653dee6038a3e951c1a19a9410c24dc393de22

No comments:

Post a Comment