Why is Backpropagation more efficient?

We already have the loss function as a function of weights.
Why don’t we just take the partial derivative using calculus?

Backprop is caculating partial derivative. Just another name of the process.

1 Like