Module 4 — Knowledge Checks (5)#
Answer in plain English. If you can explain these without equations, you understand the foundations.
Weights
What are weights, and why do we say “nothing is learned except the weights”?
Loss
What is loss, and why can’t learning happen without it?
Gradient vs Learning Rate
What does the gradient tell you?
What does the learning rate control?
What happens if the learning rate is too large or too small?
Training vs Inference
What is the difference, and why do real-world failures often show up in inference?
Overfitting & Validation
What is overfitting?
How does a validation set help detect it?
In the hiker story, what does “the scout” do?