Backpropagation and Gradients

Five questions. Choose one answer for each, then submit to see your score and explanations.

The practice target is 80%. You can retry as often as you like, and every lesson stays accessible.

You can practice as a guest. Sign in before starting an attempt to save your scores.

1. What does backpropagation compute?A. Gradients of the loss with respect to parametersB. A new dataset of correct labelsC. The best architecture without experimentsD. Only the predicted class names

2. In ordinary gradient descent, which direction is used for a small parameter update?A. Always toward a larger parameter valueB. Always in the direction of the gradientC. Toward the validation label directlyD. Opposite the gradient

3. What does a partial derivative describe?A. The average of every model weightB. A derivative that ignores all dependenciesC. How a quantity changes with one variable while others are held fixedD. Only the final output class

4. A weight is 2, its loss gradient is 3, and the learning rate is 0.1. What is the next gradient-descent weight?A. 1.9B. 1.7C. 2.3D. 0.3

5. A path has local derivatives 4 and 0.5. What is their chain-rule contribution?A. 2B. 4.5C. 8D. 0.125

Review this module from its first lesson →

Backpropagation and Gradients

Your result