Evaluate the following statements:

A perceptron $($i$.$e$.$ a single node network$)$ is not guaranteed to perfectly learn a given linearly separable function within a finite number of training steps.

Suppose a trained neural network model includes a large number of hidden nodes for a relatively simple and small$-$sized dataset. The model may show overfitting.

Consider this equation, Y $=$ ax$^2+$ bx $+$ c $($polynomial equation of degree $2).$ This equation cannot be represented by a neural network of a single hidden layer accurately.

Deep learning methods include a neural network architecture with several layers of nodes between inputs and output, using a new way to train the multi$-$layer network.

Which statements are incorrect? Choose the option below that most applies here.