Quiz · Gradient descent1 / 6
RightSignalTue, Oct 20 · today
1 of 7
Quiz · Gradient descent2 / 6
What usually happens if the step size is set far too large?
Quiz · Gradient descent3 / 6
The loss is L(w) = (w − 3)², so its slope is 2(w − 3). With w = 1 and η = 0.25, where does one step take w?
Quiz · Gradient descent4 / 6
What does the gradient tell the model at each step?
Quiz · Gradient descent5 / 6
Why might gradient descent end with a decent but not perfect answer?
Quiz · Gradient descent6 / 6
Why do models often compute the slope from a small random batch of examples?
You got 0 of 6.
6 left unanswered. Swipe back to finish them.
SwipeScroll or use arrow keys