Spam filters, credit scores and cancer screens are all one question: which side of the line?
M2 · Drawing the Line
A straight line cannot answer yes or no, and watching it fail is the fastest way to understand the fix. Bolt a soft curve on the end and you have the model that still ends almost every classifier ever built, including the ones with billions of weights. The second half is the part most courses rush: four boxes instead of one accuracy number, the two ratios everybody quotes, and a threshold chosen from what a mistake actually costs. You leave able to say what your model gets wrong, and to defend the line you drew.
- From a Line to a Label
- The Soft Switch
- Logistic Regression
- The Loss That Punishes Confidence
- Where the Numbers Came From
- Boundaries You Can See
- The Model That Never Trains
- The Four Boxes
- Caught and Cried Wolf
- Choosing a Cut From Money
- Every Threshold at Once
- When the Answer Is Almost Always No
- More Than Two Answers
- The Last Layer of Everything