Review: Two Minute Papers

Two Minute Papers is a YouTube channel that delivers what the name promises: short summaries (sometimes a bit above two minutes in length) of scientific papers, presented by Károly Zsolnai-Fehér, a young computer graphics researcher from Vienna. Given his own field of expertise, many of the featured papers are about graphics and neural networks: the simulation of realistic interaction between liquid and hair; the generation of images from sentences using a neural network; the beautification of selfies.

This has the advantage that the videos can often feature stunning animations and images from the papers that are summarized. The videos are lovingly put together, and Károly’s excitement is infectious (if occasionally a bit disproportionate). He stays away from the details and generally focuses on the practical application and the high-level challenges the researchers attempted to overcome, while giving links to any available open source code or data. There’s also a subreddit for additional links and discussion.

Like many YouTube channels, the project solicits public support via Patreon. With already more than 50,000 subscribers, the project will hopefully be able to raise enough funds to stay alive. It’s a beautiful idea, and it would be lovely to see the approach expanded to other topics. As it is, I recommend subscribing if you’re interested in graphics, machine learning, 3D printing, and similar topics.