Statistics & Math for ML · Lesson 102 of 138
Calculus Quiz
Test yourself.
Let's check your calculus. Pause before each answer.

- Five questions
- Pause and think
- Then listen
Question 1
What's the derivative of X squared? By the power rule, the answer is two X.
Quick check
d/dx of x²?
Question 2
Which way does the gradient point? The answer is uphill, the direction of steepest increase.
Quick check
The gradient points?
Question 3
What happens if the learning rate is far too big? The answer is, the steps overshoot, and training can blow up.
Quick check
Learning rate too big?
Question 4
Which rule lets backpropagation find gradients through many layers? The answer is the chain rule.
Quick check
Through many layers?
Question 5
What does stochastic gradient descent use for each step? The answer is a small random batch of the data.
Quick check
SGD uses?
Summary
Excellent. Next, the final section, the maths inside machine learning, starting with loss functions.
Key takeaways- 2x
- Gradients point uphill
- Chain rule for layers
Next up: Loss Functions — measuring mistakes