Tag: stat_ml

• Multiplexed perturbations enable massive scale ... but how big can we go? (12 May 2025)
  
• What FDR control doesn't do (20 Apr 2023)
  
• Neural Network Checklist (09 Feb 2023)
  
• Differential expression for newcomers (14 Aug 2022)
  
• You can't construct confidence intervals near a region of non-identifiability (20 Jun 2022)
  
• You can't test for independence conditional on a quantitative random variable. (20 Jun 2022)
  
• You can't estimate the mean. (20 Jun 2022)
  
• You can't satisfy three simple desiderata for a clustering algorithm. (20 Jun 2022)
  
• You can't achieve optimal inference and optimal prediction simultaneously. (20 Jun 2022)
  
• Five disturbing impossibility theorems (20 Jun 2022)
  
• A third miracle of modern frequentist statistics (15 Jun 2022)
  
• A second miracle of modern frequentist statistics (15 Jun 2022)
  
• A miracle of modern frequentist statistics (15 Jun 2022)
  
• A fourth miracle of modern frequentist statistics (15 Jun 2022)
  
• Lazy matrix evaluation saves RAM in common analyses (28 Nov 2020)
  
• Kalman Filter Cheat Sheet (30 Aug 2020)
  
• Experiments with one control and $k$ treatments have highest power when the control arm is $\sqrt k$ times bigger than each individual treatment arm (21 Aug 2020)
  
• Favorite machine learning papers (resnets) (21 Jul 2019)
  
• I don't want to be afraid of neural networks anymore (06 Jul 2019)
  
• We want R functions for gradients and hessians (12 Jan 2019)