Michael Ekstrand

One Tweet Research

Published on 21 Mar 2017 and tagged with research and communication.

I spent the day yesterday at the NSF CISE CAREER workshop. It was excellent, and I strongly encourage any young US CS faculty to go to a future installment or watch the online videos if a CAREER proposal is in their future.

One thing that came up repeatedly is the need to clearly — and concisely — explain your research. Several program directors referenced the Heilmeier Catechism, the first question of which is “What are you trying to do? Articulate your objectives using absolutely no jargon.” Another brought up that it's important to be able to explain briefly what you're doing (or planning to do).

I've been adopting something like this in my talks for a few years now. When I'm giving a research seminar, one of my first few slides is ‘#1TweetResearch’: how would I convey, in a single tweet, the overall thrust of my work? Here's the slide from my last seminar:

#1TweetResearch: How can we make recommender systems good for the people they affect?

I started this practice with my thesis defense, as you can see here.

In individual research paper talks, I do something similar, but with a different frame: I write a 'TL;DR' slide early with the high-level takeaways, like this one.

I also try to do this in my writing. In these blog posts, as you can see, I include a brief “TL;DR” summary at the beginning. I try to make my paper abstracts similarly informative.

I also have no qualms whatsoever about revealing the conclusion in the summary. This is science, not a mystery novel. Spoilers are entirely fine. As a reader (and particularly a reviewer), I appreciate them quite a lot. If I already know where the paper (or talk) is going, then all I have to ask myself while reading it is whether the author is convincing me of the point. I don't have to simultaneously try to figure out what that point is.