Cochin University of Science and Techology (CUST) 2006-7th Sem B.E Computer Science and Engineering B.TECH , - Question Paper

SUB:-ARTIFICIAL NEURAL NETWORK,JUNE 2006
COLLEGE:-SOE, CUSAT
BRANCH:-EC,CS,EB

B. Tech Degree VIISemester (Supplementary) Examination

June 2006

EB/EC/CS 705 (C) ARTIFICIAL NEURAL NUT WORKS

0-rior to 2002 A dm issions)

Time : 3 Hours    Maximum Marks : 100

L (s) Design a MLP network t solve the two input XOR problem.    (12)

(b) Differentiate between supervised and un-supervised leinsng algorithms with examples.(8)

OR

H. (a) What is ddta learning id e?Ust the iintftations.    (8)

(b) Witte down the expression for sigmoidal and Gaussian tctivaiion (Unctions and draw them. How activation functions are classified?    (12)

HI. (a) Differentiate between learning by epoch and learning bj pattern.    (8)

(b) Explain the BP algorithm with momentum and smoothii ig? How it improves the

performance of standard HP algorithm?    (12)

OR

IV. (a) What do you mean by ovar learning? How the problem i s eliminated?    (8)

(b) How the local minima pniblem is dealt in BP training. Mention two stopping criteria in BP algorithm.    (12)

V.    (a) How training is applied tc Kohcnen layer? Explain with relevant equations.    (10) (b) Explain one typical application of counter propagation n eural networks.    (10)

OR

VI.    (a) What is preprocessing of :input vectors? Why it is require*!? How initial weights are

decided in a network? How these factors will affect the network performance?    (12)

(b) Explain the structure of a feed forward counter propagatj on network.    (8)

VIL (a) What is simulated annealiag? Explain how the concept ii applied in network training? (10) (b) Explain how neural netwerk is allied in optimization pi oblems.    (10)

OR

VEOL (a) Compare the features ofEP and Cauchy's training algorithms.    (10)

(b) Write down the Boltzman training algorithm and explain the steps.    (10)

IX.    (a) How discrete Hopfield network stores binary patterns? Explain the capacity

limitations.    (10)

(b) Explain mutation and crossover operations in GA with ex amples.    (10)

OR

X.    (a) How the data retrieve operation is performed in BAM?    (10) (b) Explain the stroctureof AJT-I type networks.    (10)

Attachment: cochin-university-of-science-and-techology--cust--2006-b-e-computer-science-and-engineering-36712-1.pdf

Earning:   Approval pending.