$1.$ Define an Artificial Neuron Network $($ANN$)$ model as;

$-5$ input, $2$ hidden layer with $4$ neurons each and $2$ outputs. The weight coefficients will be chosen randomly $\backslash ([0,1]\backslash ).$

$-$ For activation functions: linear, threshold, sigmoid nonlinear cases.

First draw the ANN structure and then apply $3$ different activation functions for each one. Then curve outputs of the ANN structure.

$2.$ Give $3$ practical example for the pseudo$-$code given below;

Activation functions for each one

function Simple$-$Problem$-$SOLVING$-$AGENT $($ percept$)$ returns an action persistent: seq, an action sequence, initially empty state, some description of the current world state goal, a goal, initially null problem, a problem formulation

state $\backslash (\backslash$leftarrow $\backslash )$ UPDATE$-$STATE $\backslash ((\backslash )$ state, percept$)$

if $\backslash ($ s e q $\backslash )$ is empty then

goal $\backslash (\backslash$leftarrow $\backslash )$ FORMULATE$-$GOAL$($state$)$

problem $\backslash (\backslash$leftarrow $\backslash )$ FORMULATE$-$PROBLEM$$state$,$ goal$)$

seq $\backslash (\backslash$leftarrow $\backslash$operatorname$\{$SEARCH$\}(\backslash )$ problem $\backslash ()\backslash )$

if $\backslash ($ s e q$=\backslash )$ failure then return a null action

action $\backslash (\backslash$leftarrow $\backslash$operatorname$\{$FIRST$\}($s e q$)\backslash )$

$\backslash ($ s e q $\backslash$leftarrow $\backslash$operatorname$\{$REST$\}($s e q$)\backslash )$

return action

$3.$ Write $3$ different sample programs for pseudocode agent programs for the goal$-$based and utility$-$based agents.