Question $1[2$ pts$]$: Figure $1$ shows some samples in a two$-$dimensional feature space $($X$1,$ X$2).$

Each symbol $($circle$,$ triangle, plus, diamond$)$ denotes a sample, and each shape denote one type

of samples. Please design a neural network architecture which can learn to classify samples in

Figure $3$ into correct types.

$$ Show your network input dimension, number of hidden nodes $($layers$),$ and output

node$($s$)[1$ pt$].$

$$ Explain why you use this network architecture $[1$ pt$]$

Figure $1$