Quality First and Other Advice
Tags: academia, musings
At our local meetup for the ‘Learning on Graphs’ Conference , we also organised a brief panel discussion with the idea of engaging researchers at all levels. While every panellist had a different story to tell, there is nevertheless one common strategy that I extracted from their discussions: When in doubt, focus on quality over quantity.
I can echo this sentiment. The few times in my research career where I lost sight of the quality of a publication, I paid for it afterwards. Shoddy work—even if it gets accepted—will always come back and haunt you, forcing you to correct your mistakes afterwards. I know that in the current age of machine learning research, there are students graduating with lots and of papers, but bear in mind that the number of papers, if anything, is also a measure of the lab size: Bigger labs tend to make it easier for Ph.D. students to participate in many projects. The numbers may thus be highly deceptive.
Notice that I am not advocating for complacency or never submitting anything to a conference during your Ph.D. On the contrary, I believe that navigating the ambiguous territory between ‘This is not ready yet,’ ‘This is good enough to get useful community feedback, and ‘This is perfect’ is one of the core skills you have to acquire as a researcher. Perfection is the enemy of getting things done, but low standards will make you miserable over the long run. Thus, as in all things, balance is key.
I also believe that focusing on quality enables you to find more joy in your research as you hone your craft. Since a Ph.D., and research in general, can sometimes be extremely frustrating, it is important that you only do the work you appreciate. Do not publish in a way that, afterwards, makes you feel empty, because, again, you will pay a price for shoddy work.
Somewhat related to low-quality research: Resist the urge to pull all-nighters. It seems like a good idea at the time, but I can attest to the fact that none of the papers we pushed very hard for during the night ever saw the light of the day—at least not without many additional major changes. The only thing our all-nighter accomplished was decreasing our motivation to work on the project. Add to this the high likelihood of getting strongly-worded reviews, and you create something that can sap motivation and life even out of the strongest people. Good research needs time and learning when the sunk cost fallacy has you in your grips is also something I believe will come in handy not only in academia.
Until next time, I wish you all the best in carving your own path!