How does Gradient Descent help in training a Deep Neural Network $($DNN$)?$

Group of answer choices

It decreases the complexity of the model by removing unnecessary layers and nodes.

It decreases the error rate $($or loss function$)$ of the model by changing its parameters iteratively.

It increases the error rate $($or loss function$)$ of the model by changing its parameters randomly.

It increases the complexity of the model by adding more layers and nodes.