Question 8
Domain 2 — Data, Machine Learning, and Model DevelopmentIn deep learning, what is "backpropagation" used for?
Correct answer: B
Explanation
Backpropagation is the training method used to compute how much each weight contributed to the output error and then send that error backward through the network. This lets the model update weights to reduce future error, which matches "calculating errors and updating weights."
Why each option is right or wrong
A. Initializing weights in the model
B. Calculating errors and updating weights
Backpropagation is the standard training procedure for neural networks: it applies the chain rule to compute the gradient of the loss function with respect to each weight, then uses those gradients to adjust the weights during optimization. In practice, the error signal is propagated from the output layer back through earlier layers so the model can reduce the loss on the next update step.
C. Organizing data into categories
D. Reducing the model's complexity