Academic Origami

Turning scientific publications into paper cranes

Origami is the art of transforming a piece of paper into a beautiful object, or, as often as not, into a crumbled ball in the basket. That’s what I intend to do with this new series of posts on scientific papers I read. Either share my understanding and help everyone marvel at an arcane piece of science, or blatantly expose my poor comprehension for everyone to see.

My hope in doing this is threefold:

1. Organize my own thoughts on a scientific paper,
2. Help others who might be struggling as much as I am,
3. Appeal to anyone who understands the subject and wants to help me.

If I’m honest there’s a fourth objective: encourage others to do the same so that I can read digested versions of papers in the future.

How does it work?

An origamized paper will be made from excerpts of the original paper (in the form of image captures) interspersed with my informal explanation, my interrogations and my opinions. The best way to read a post is top-to-bottom (no kidding!) then whenever you encounter an excerpt from the original paper, read it first before you read the text beside it.

Each paper I origamize will be published on Medium. Everyone is welcome to comment right on the published document using Medium’s highlight-and-comment tool. I’ll fold interesting recommendations into the main text, or at the very least answer them.

Principles

Any such grand undertaking needs a set of pretentious principles, here are mine:

• Focus on ideas, not people.
• Be casual and clear, not concise and formal.
• For previous work, see previous work.
• For the full paper, see the full paper.
• Opinions help.

Corollaries

I’ll refer to papers by what they talk about, not their authors. I know, I’m breaking an age-old academic tradition. Deal with it. :)

I’ll use the same casual language I use everyday in emails or on Facebook. This is how I like to write and, incidentally, that’s the kind of stuff I like to read. Yes, there will be smileys.

I will not cover previous work. I will aggressively de-duplicate information. I will aggressively cut parts of the papers I do not consider helpful, or which I prefer to rephrase. I will very rarely include results.

Last but not least, I will not shy away from voicing my opinion. The goal is not to discredit ideas, but rather to help readers build a better mental model of the field. Which idea had a critical impact on later papers? Which idea was a dead end? Which one sounds promising? What part of the paper seems impossible to understand? I’ll basically be offering answers to the questions I always have whenever I read a paper.

Keep in mind

1. I’m going to get things wrong.
2. I’m expressing an opinion on an idea, not on yourself. Also, see (1).
3. I’m not an expert. If I were, I wouldn’t be reading these papers, so (1).

Therefore please correct me. Always. Counter my opinion with your own. If you make a good case for it you’ll likely succeed in changing what I think. This will make me happy and the world will sing.

Conclusion

It’s an experiment. I’ll see how long it lasts. I’m reading a bunch of papers these days and I find that keeping a Medium post on the side isn’t too hard and brings me much benefits. I expect to keep doing it for a while. However if you think what I’m doing brings value don’t hesitate to let me know. A tap on the back works wonders for me. :)

And now, time to read some papers!