|
Register Number
|
|
|
|
|
|
|
|
(Established under
section 3 of UGC Act,1956)
Course & Branch :B.E - P-EEE
Title of the Paper :Fundamentals of Fuzzy Logic and
Artificial Neural Networks Max.
Marks :80
Sub. Code :SECX1048(2010/2011) Time : 3 Hours
Date :20/03/2014 Session
:FN
______________________________________________________________________________________________________________________
PART -
A (10 x 2 = 20)
Answer ALL the Questions
1. State the characteristics of a linguistic
variable.
2. Define cross over points of a fuzzy set.
3. Write the fuzzy rules for solving XOR
problem.
4. If
universe of integers is y={1,2,3,4,9}, the following linguistic terms are
mapped on to Y.
“Small” = 
“Large”=
Find the composite linguistic term = “not very small
and not very large”
5. What do synchronous and asynchronous updates
mean?
6. Construct
McCulloch Pitts network for the following expression.
N3(
t ) = N2( t – 2 ) & N1( t- 3 )
7. Compare the performance of Boltzmann and
Cauchy machines
8. Why
convergence is not guaranteed in back propagation networks?
9. Prove
that BAM is unconditionally stable for any binary units.
10. Distinguish
complex cells from simple cells in a neocognitron.
PART – B (5
x 12 = 60)
Answer All the Questions
11. (a)
Let two fuzzy sets be given as (4)
A=
and B= 
and B=
Verify
Demorgam’s principles.
(b)
Let two fuzzy relations be given as (8)
R1=
and R2=
and R2=
Find
R1 Ο R2 by i.
Max-Min composition
ii. Mazx-product composition
(or)
12. For
the following output fuzzy sets B1, B2 and B3, calculate the defuzzified value
using
(a) Centroid method
(b)
Weighted Average
(c) First/Last of Maxima
13. With
suitable example explain the applicability of fuzzy logic in measurement,
control and manufacturing fields.
(or)
14. Discuss
about the application of fuzzy logic in Biomedical signal processing.
15. (a)
With neat sketches explain the structure and function of a typical cell in the
central nervous system. (8)
(b)
Compare Biological Neural Network with Artificial Neural Network. (4)
(or)
16. (a) Give a simple perceptron neuron model and
explain why perceptron cannot represent an EX-OR problem?
(b) Find the updated weights of a
perceptron model to solve NAND problem. (Assume Bipolar input)
17. Analyse
a single feed forward and back propagation step for the network shown It should
respond with [O1 O2] T = [ 0.95
0.05] T to the input pattern [Z1 Z2 -1]
T = [ 1 3 -1] T
(or)
18. (a)
Describe the Instar and Outstar model and explain how counter propagation
network can be used for image classification.
(8)
(b)
Assume the three memories of five dimensions as follows are to be stored in a
Hopfield model. Find the weight structure.
(4)
I1
= { 1, 1, 1, 1, 1 }, I2 = {1, -1, -1, 1, -1 } and I3={
-1, 1, -1, -1, -1 }
19. What
are cognitron and Neocognitron? With suitable diagram explain the S-cell and
C-cell processing algorithm of Neocognitron.
(or)
20. Consider
an ART network with nine input units and two cluster units. After some training
the bottom-up weights (bij) and top-down weights (tji)
have the following values.
|
Bottom up weights
|
|
Top-down weights
|
|
|
0.4
|
0.25
|
1 0
0 1 1
0 0 1 1
|
|
|
0
|
0.25
|
1 1
1 1 1
1 1 1 1
|
|
|
0.4
|
0.25
|
|
|
|
0
|
0.25
|
|
|
|
0.4
|
0.25
|
|
|
|
0
|
0.25
|
|
|
|
0.4
|
0.25
|
|
|
|
0
|
0.25
|
|
|
|
0.4
|
0.25
|
|
|
The
pattern (1, 0, 1, 0, 1, 1, 1, 1, 1) is presented to the network. Compute the
action of the network if the vigilance parameter is 0.5.
0 comments:
Pen down your valuable important comments below