CUSAT Question paper Artificial Neural Networks Nov 2006

CUSAT Question paper Artificial Neural Networks Nov 2006

CUSAT B.Tech. Degree VII Semester Examination, Nov 2006

 

IT/CS/EC/EB/EI 70S (C) ARTIFICIAL NEURAL NETWORKS

(2002 Admissions onwards)

Time:

3 Hours

Maximum Marks: 100

I

a)

State and explain the perceptron learning algorithm.

(8)

b)

What is linear seperability? Why can’t the single layer perceptron implement

an X-OR gate? Explain.

(12)

OR

II

a)

With relevant equations explain Hebbian learning rule.

(8)

b)

Compare LMS, perceptron and delta learning rules.

(12)

III

a)

Explain how momentum method improve the training time of the back propagation

algorithm.

(10)

b)

Explain the outer product rule.

(5)

c)

When the training in a network stopped?

(5)

. OR

IV

a)

Write short notes on storage capacity.

(6)

b)

Explain stability and convergence in connection with ANNs.

(6)

<0

Explain the relevance of the learning rate parameter rj in B.P. algorithm. How it will

affect the learning process?

(8)

V

a)

Explain the Kohonen’s map of self organizing neural networks.

(10)

b)

Discuss the use of Kohonen’s model in feature extraction applications.

(10)

OR

VI

a)

Explain how counter propagation networks differs from feed forward networks.

(10)

b)

In connection with training, explain the concepts of pre-processing of input vectors

and initialization of the weight vectors.

(10)

VII

a)

How a Boltzman machine works? Mention two applications.

(10)

b)

Explain anyone non-linear optimization problem where neural networks can be

employed.

(10)

OR

VIII

a)

Explain simulated annealing in detail.

(10)

b)

How Boltzman training is different from Cauchy training.

(10)

IX

a)

What are the characteristics of ART?

(10)

b)

How data is stored and retrived in BAM?

(10)

OR

X

a)

Explain mutation and cross over in genetic algorithm. What are the applications of

genetic algorithm?

(10)

b)

Explain the stability condition in a Hopfield network. ^-

(10)

***

 

 

 

Leave a Comment