paraphrasing one of the greats: it's already there, just chisel away the unnecessary bits

Join millions who have switched to Grok.

shocking insights, i know.

i used twitter a lot recently, sometimes multiple hours per day . my two biggest takeaways:.- harm reduction on social networks is easier when you know about the darkish art of rhetoric.- you interact (mostly) with humans. Few suck, but others do not. Just don't engage the former.

RT @WillShenSaysHi: Real dog meets robot dog at CoRL! #unitree #CoRL2022

big thanks to @albertopessia for invaluable advice during the model development, @NetPharMed for the support and patience, and the clinical collaborators for asnwering all my silly questions. Last and def not the least, we'd be nowhere without patients and their consent!.

4/n learn to love your data <3.

3/n try to have Cox regression as a baseline and give a chance to good-old decision trees, e.g optimal survival trees by Bertsimas etal.

2/n your prognostic models would have a higher chance of being using in the clinic, if they are not black boxes.

we published a paper! 1/n if you wanna do survival analysis, make sure to assess both model discrimination and calibration, using eg C-index and Brier score. Especially if you are using fancier ML methods.

RT @philipvollet: NN-SVG is a tool for creating Neural Network architecture drawings parametrically rather than manually!. It also provides….

so nice you can solve any overfitting by ensembling a few dozen distinct models.

RT @f2harrell: #Statistics thought of the day: If you must use any cutpoints for continuous variables, only use them for the 3rd derivative….

RT @ChristophMolnar: Machine learning sucks at uncertainty quantification. But there is a solution that almost sounds too good to be true:….

RT @future__art: Hey guys I fixed another Monet painting

that first paragraph

RT @slavov_n: Let’s talk about proteins whose abundance is generally independent from the abundance of the transcripts that templates them.….

RT @LabKauppi: Elli's article on Differential Insertion-Deletion Dynamics of Microsatellites in Mouse T Cells using single-cell exome seque….

RT @nickcammarata: What specific scientific paper you can link to most contradicted your prior belief and made you 180° to a new one.

also according to Barabási's tweet philanthropic science funding exceeded that of NIH and NSF, both in amount and number of grants. I am just gonna assume that people are good and no oversight is needed on those decisions.