During the backpropagation algorithm, what role does the chain rule play?

Group of answer choices

The chain rule helps in calculating the gradient of the loss function with respect to the network's output.

The chain rule helps in determining the optimal learning rate for updating the network parameters during training.

The chain rule is used to initialize the network parameters before training begins, ensuring convergence to a suitable solution.

The chain rule is used to propagate errors backward through the network by computing the gradient of the loss function with respect to each layer's inputs.