Lesson 0001
The one diagram that everything else in this course hangs off. If you remember only one thing from Lecture 3, remember this.
STOR 323 defines machine learning operationally: performance on a task T improves through experience E, measured by performance P. For MNIST digit recognition: T = recognize a handwritten digit, E = labeled examples, P = error on new digits. Nothing mystical — just data in, better performance out.
The whole course splits into two phases that are easy to blur together if you don't separate them explicitly. Training happens once, ahead of time. Deployment happens over and over, on new inputs.
Training: a fixed learning algorithm A reads a dataset D and returns a fitted model f̂D.
Deployment: a new input x is predicted into a score s, then decided by a rule δ into an action a.
Every later topic in this deck is really a question about one arrow in this diagram:
| Question | Which arrow |
|---|---|
| Which functions could f̂D possibly be? (linear regression, softmax regression, kNN) | D --A--> f̂_D — the model family constrains A |
| How good is f̂D's prediction s? (MSE, ROC/AUC, confusion matrix) | x --f̂_D--> s |
| What action should a score trigger, and who bears the cost of a wrong one? | s --δ--> a |
The deck's punchline (slide 21): the model is fixed; the policy is not. You can swap δ — change a threshold, change who gets referred for a test — without retraining anything. That's a deployment-time decision, not a modeling one. Future lessons will dig into model families and evaluation metrics separately; this lesson is just about being able to place any concept on this diagram without hesitation.
Answer from memory before checking — that effort is what makes it stick.
Primary source for this lecture's formalism: An Introduction to Statistical Learning (James, Witten, Hastie, Tibshirani) — free official PDF, Ch. 2 covers exactly this train/predict framing. See RESOURCES.md for more.