Egor Revkov

PUREE is a weakly supervised machine learning algorithm that can accurately infer tumor purity from bulk tumor gene expression data: doi.org/10.1038/s42003… @anders_js, @RevkovEgor

⚡️Preprint: Why do tree-based models still outperform deep learning on tabular data? We give solid evidence that, on tabular data, achieving good prediction is easier with tree methods than deep learning (even modern architectures) and explore why hal.archives-ouvertes.fr/hal-03723551 1/9

When you realize that collecting big data is great but in the end you have to chase down the genes and understand how they work

Let's fix academic posters! Prepping a poster for #SIOP19 and sick of the old "wall-of-text" poster design? Watch this cartoon to see a new, faster approach to designing research posters. Includes templates. #betterposter youtu.be/1RwJbhkCA58

As a budding student of ML, I often find myself re-googling things I've learned/forgotten many times. This afternoon, I decided to toss some favorite resources into one doc for speedy reference. Making it for my own use, but figured why not share the link sgfin.github.io/learning-resou…

"Bag of Tricks for Image Classification with Convolutional Neural Networks" is such a great paper. Really clear, and most importantly, amazing results - >94.5% imagenet with a simple rn50, beating even inception v3! I'm working on replicating this now with fastai.

"to be explicit about how new statements modify the plot; repetition should facilitate more interactions with ggplot functions, helping newcomers internalize these functions. The flipbook layout is designed to show what code is triggering each change in the plot output."

Amazing latex plots for CNNs github.com/HarisIqbal88/P…

My institution has been hiring a ton, which means I've attended over 20 job talks in the last 12 months. It occurred to me a thread about DOs and DON'Ts might be helpful, so here goes:

many years of git in, i've developed lots of muscle memory for working around git disasters, but this would have been helpful to me when i was starting out ohshitgit.com

Identification of somatic mutations in cancer NGS data is NOT a solved problem. Our new ensemble method is just out with much improved accuracy over existing algorithms: dx.doi.org/10.1093/bioinf…