pase $7$ of $8$

ITEC$460$ introduction to Newral Networks

Final Exam $-$ a

A network should be designed to store $10$ pattern$-$associations. A sample association is shown below. Which of the following neural network architecture can be used for this problem?

a sample pattern association

a$)$ A $10810$ multi$-$layer perceptron

b$)$ A $2020$ heteroassociative memory neural network

c$)$ A $2010$ radial basis function neural network

$-$d$)$ A $2020$ autoassociative memory neural network

How many adjustable parameters $($weights$)$ does a Hopfield network with $8$ units have?

a$)8$

b$)16$

c$)56$

d$)64$

Which of the following artificial neural networks uses unsupervised learning?

a$)$ Multi$-$layer perceptron network

b$)$ Radial basis function network

c$)$ Hopfield network

d$)$ None of them

Which of the following statements is NOT true for a RBF networks?

a$)$ RBF networks can be used for approximating functions.

b$)$ The weights of the second layer are updated by the delta rule.

c$)$ The number of hidden units usually exceeds the number of input units.

d$)$ RBF networks are examples of nonlinear layered feed$-$forward networks.

e$)$ The sigmoid function is the most commonly used Radial Basis Function.

What is an associative memory?

a$)$ A neural network where the desired outputs are equal to the inputs and a set of patterns are stored in the hidden layer.

b$)$ A neural network which can store a set of patterns, and recall the nearest stored pattern when an incomplete pattern is presented to the network.

c$)$ A neural network which can solve classification problems by learning to associate pattern vectors with the desired outputs using backpropagation rule.

d$)$ A neural network which can solve linearly inseparable pattern recognition problems such as face recognition.

$7$