Paper a Week

Paper a Week began as an experiment in personal accountability when I was a 3rd year PhD student. The idea is simple: read and take structured notes on one academic paper per week. Meeting that goal has helped me build a literature reading habit and organize my notes in a searchable, sortable, and shareable format. Wanting to share the benefits of this system with others, I built out Paper a Week into a standalone service capable of finding papers online, maintaining a reading list, writing rich notes with Markdown and LaTeX support, and providing habit-tracker statistics. Making an account is fast and free. Try it out!

Visit the Paper a Week Website

Explore the Code

A big part of my motivation for making Paper a Week was to learn modern web technologies including React, Redux, Express, Node.js, MongoDB, Docker, Google Cloud Run, Typescript, and more. That effort would have been much more painful if not for the expert guidance and patience of my mentor and identical twin brother, Arad. Given the nature of the project, I've made all of the code and commit history available to anybody interested.

Reflections

As of this writing, I had done 200 consecutive weekly reviews. My 200th review aligned with finishing my PhD, and I took the opportunity to take a break from weekly reviews and focus on other things. Here are my thoughts on Paper a Week after nearly 4 years of regular use:

• On Frequency: A lot of people might see 1 paper / week as unambitious, but I’ve found that it’s a great balance of regularity and sustainability. There were times in my PhD (like before my qualifying exam) where I was carefully reading many papers in a week, but there were also times — whole quarters sometimes! — where other things felt more urgent and I barely skimmed any papers, let alone read them carefully. 1 paper per week is a tangible target, and if I ever wanted to read more than that, I did.
• On Skimming: Taking notes on 1 paper / week doesn’t necessarily mean looking at 1 paper / week. For every paper I took notes on, there are 3-5 that I skim or stop reading halfway through. In fact, by choosing “The One” paper I’ll read that week, I usually end up skimming many alternatives and reckoning with whether a paper is actually relevant to my interests.
• On Old and New Literature: Reading a paper-a-week helped me do two things that I find important: First, it pushed me to actually scan the new literature on a weekly basis, instead of relying on Twitter to float a relevant article once a month. Second, in weeks where there was nothing relevant and new to read, it pushed me to find older papers that had flown under my radar.
• On a Personal Scientific Brand: Having a list of many papers that were of interest to me has helped defined my messy corner of scientific interests. Reflecting on all of the notes I’ve taken, the intersection of my interests in machine learning, vision, and cortical development is more apparent. After many weeks, Paper a Week provides an honest reflection of the work you find interesting, accessible, and important.
• On Memory: My memory for papers I’ve read isn’t great, and I’ve already used paperaweek.com to go back and remind myself of what my impressions were at the time I read a paper.
• On Notes as a Form of Reviewing: Taking critical notes on one paper per week has helped make me a better reviewer and critic of scientific work in journals, talks, and conferences. It's not a replacement for peer review, but it's still made me better at noticing details in papers.