🔥Telegram İfşa

Introducing ML Visuals, a collaborative effort to help the ML community improve science communication by providing free professional & compelling visuals. Feel free to use them in your own presentations/blogs. Visuals: docs.google.com/presentation/d… Read more: github.com/dair-ai/ml-vis…

For my optimization theory homies...

Integrative computational models can leverage genomic data to understand the impact of genetic variation, the effect of dysregulated genes on biological pathways in specific disease and tissue contexts, and disease risk

They highlight existing methods and key challenges and opportunities in identifying specific disease-causing genetic variants and linking them to molecular pathways and, ultimately, to disease phenotypes

Excited to share a recent review from Troyanskaya Lab - check it out!

A bunch of new video lectures from our Deep Learning course (Spring 2021 edition) put together by Alfredo Canziani

When predicting risk scores for a (primary) disease using #PRS, why not consider the effects of other comorbidity diseases (secondary/complications)? Check out our netCRS paper! #PSB22 Dept. of Biostatistics, Epidemiology & Informatics Penn IBI psb.stanford.edu/psb-online/pro…

[6/6] Link to the paper: "Why do tree-based models still outperform deep learning on tabular data?" arxiv.org/abs/2207.08815

Writing is an absolutely critical skill… and one that can be improved dramatically. Doing so requires not just repetitions, but deliberate efforts to improve. Some tips here: docs.google.com/document/d/1jB…

When discussing the relationship between Transformers and GNNs, a common pitfall is the softmax normalisation. Here's a brief writeup showing how this issue can be overcome by clever bookkeeping... and arguing why I don't expect a simple expressive GNN to robustly learn this.

📢 Excited to release PyG (PyTorch Geometric) v2.5, including distributed training, a dedicated graph tensor representation, RecSys support, PyTorch 2.2 and native compilation support 🎉 Release Notes 👇 github.com/pyg-team/pytor… 🧵 A thread [1/6]