Anaibram in

Pacedefs

Frcese introhktion to Neural Nenveris

Which of the following statements is NOT true for a Hopfield network?

a$)$ The weight matrix is symmetric:

b$)$ A unit can be connected to itself.

c$)$ The units in a Hopfield network can be updated randomly during recall.

$-$d$)$ The output of the units is often specified by a bipolar step $($sign$)$ function.

Which of the following statements is true for the centers of the Gaussian activation functions of radial basis function networks?

a$)$ They define the first layer of weights.

b$)$ They are the values of the activation functions.

c$)$ They define the width of the activation functions.

d$)$ They are the prototype vectors.

What is the minimum number of units $(M)$ necessary to store and recall six patterns successfully by a Hopfield neural network? The theoretical limit for the maximum number of patterns that can be stored is given by $\frac{N}{2lnN},$ where $N$ is the number of network units.

a$)2$

b$)18$

c$)36$

$($d$)46$

e$)60$

Which of the following RBF network architectures would be most appropriate to predict "Blood Pressure" based on four features $("$Gender$,"$ "Age," "Heart Rate," and "Body Temperature"$)?$

d$)$ A $141$ RBF network

b$)$ A $421$ RBF network

c$)$ A $461$ RBF network

d$)$ A $144$ RBF network

e$)$ A $464$ RBF network

$4$