HN Explorer

Ask HN: How to get start with publishing research papers?

5 points by vs2370 ↗ HN
I work as an engineer in machine learning/big data field. Its not part of my day job but I wanted to get started with publishing research papers. A research portfolio will help me with job, learning, immigration and higher studies.

There is no one I can find to pair with. I am willing to do a lot of engineering work, data work but I have no experience when it comes to writing and publishing papers. A lot of papers from ICML and NIPS are beyond my understanding even though I can easily implement classification, regression, etc.

Is there a platform or community that could be of help?

3 comments

[ 0.21 ms ] story [ 14.5 ms ] thread
deepnotderp
Try arxiv, although I'll warn that if ICML and NIPS papers are too complicated for you, then you should probably stick to application papers like lipnet.
probably_wrong
I'm afraid the only community I know which teaches how to write papers is typically your local University. If that's not an option (understandably), at the very least you want to partner up with someone who did it before.

Look for coding meetups in your area, and find someone with experience. It's not unusual to work in pairs, where the more experienced one writes the paper while the other one runs the experiments.

If you are having troubles reading papers, that's your first step: sit down with simple, short papers, and work through them. Reading papers, just like reading code, it's a skill that you obtain with practice.

gus_massa
I agree with this. There is a lot of unwritten protocol in publishing and it's usually better to pair with someone that know how to select the journal, fight with the editor and referees, how many cites to use, ...

But be careful because there are also workaholics in academia, so select someone that is happy with your amount of time dedicated to this.