A neural network is a multiple$-$step function for taking some input and predicting a corresponding output. To make accurate predictions, a neural network must be "trained". What does training consist of$?$

Question $2$ options:

Learning a set of weight parameters by looking at many training examples.

Reconfiguring the number of nodes and layers until a minimum accuracy is achieved.

Copying parameters from other neural networks.

Incorporating a set of hand$-$written rules from a programmer.