You are on page 1of 2

Roll No.

KRISHNA ENGINEERING COLLEGE, GHAZIABAD


Department of CSE-AI/ CSE-AI-ML
B. Tech. (SEM: V)
CT (ODD Sem-2023-24), October 2023
Subject Name: Application of Soft Computing Subject Code: KCS056
Max. Time: 2Hrs Max. Marks: 50
Faculty Name: Dr. Pratibha Singh

Identify and describe soft computing techniques and their roles in building intelligent machines and
CO 1
signify the relation between real brains and simple artificial neural network models.
Identify and apply learning methods of artificial neural networks (as perceptron and back propagation networks) in
CO 2
real world classification problems
Apply fuzzy logic and reasoning to handle uncertainty and solving interdisciplinary engineering problems.
CO 3
Note: Attempt all sections.

SECTION – A
Attempt all questions. (2x4=8)
Q. No. Question Marks CO
Compare and contrast biological neuron and artificial neuron. Draw diagrams of 2 1
a a biological neuron and the equivalent artificial neuron, with the key
components labeled.
Suggest which type of learning algorithm you will use for following cases with 2 1
justification:
1. b a. Planning strategies in chess playing
b. Classification of spam in emails
c. Document clustering
What is the effect of learning rate on the convergence of the back-propagation 2 2
c
learning algorithm
d Compare crisp relation with fuzzy relation. Give examples. 2 3

SECTION – B
Attempt all questions. (6x3=18)
Q. No. Question Marks CO
A) Explain McCulloch-Pitts Neuron Model and write disadvantages of it.
Design networks of McCulloch-Pitts neurons that implement logical OR gate.
OR
2. B) What is an activation function? What is the necessity of activation function? 6 1
List the commonly used activation functions. If the net input to an output
neuron is 0.32, calculate its output when the activation function is bipolar
sigmoidal.
A) What is the learning rule of Rosenblatt’s Perceptron model? Consider the
training set given below:
No. Input Values Desired Output Values
1 (1,0,1) -1
2 (0, 1,-1) 1
3. 6 2
3 (-1,-1, -1) 1
Assume learning constant to be 0.1 and the initial weight vector is w0= (1; 1; 0).
Use Signum as an activation function for generating output. Illustrate the
process of perceptron learning for at least three iterations.
OR
B) What are linear and non-linear separable problems? State the drawbacks of
single layer neural networks. Name a problem which cannot be solved by the
single layer neural network.
A) What do you understand by min-max composition of two fuzzy relations?
Let X = {x1, x2, x3}, Y= {y1, y2}, Z= {z1, z2, z3}. Determine the matrix for 𝑅̃ ●𝑆̃
by max-min composition.
Let 𝑅̃ be a fuzzy relation: y1 y2
x1 0.3 0.4
x2 0.2 0.8
x3 0.7 0.5
Let 𝑆̃ be a fuzzy relation:
z1 z2 z3
4. 6 3
y1 0.5 0.3 0.6
y2 0.4 0.7 0.8
OR
B) The task is to recognize English alphabetical characters (F, E, I, T) in an
image processing system. Two fuzzy sets 𝐼̃ and 𝐹̃ are defined to represent the
identification of characters I and F such that 𝐼̃= {(F,0.4),(E,0.3),(I,0.9),(T,0.8)}
and 𝐹̃ = {(F,0.99), (E,0.8),(I,0.5),(T,0.5)}. Find the following (i) 𝐼̃ ∪ 𝐹̃ (ii) (𝐼̃-𝐹̃ )
(iii) 𝐹̃ U 𝐹̃ c (iv) Verify De Morgan’s law

SECTION – C
Attempt all questions. (8x3=24)
Q. No. Question Marks CO
A) Explaining Hebb rule of learning, find the weights required to perform the
classification between input vectors and the output vectors. Input Vectors are: (1 -1 1 -
1) and (1 1 1 -1) belong to output class (target value +1) and (-1 -1 1 1) and (1 1 -1 -
5. 1) belong to output class (target value -1). 8 1
OR
B) What is neural network architecture? Explain the different types of neural networks.
Briefly discuss the common application domains of an artificial neural network.
A) What is backpropagation learning? Derive the weight updation formulae between
the output layer and hidden layer.
OR
B) What is the difference between auto and hetero correlators?
For the given input vectors S = (S1, S2, S3, S4) and output vectors t = (t1, t2), find the
weight matrix using a hetero-associative training algorithm.
6. 8 2
S S1 S2 S3 S4 t1 t2
I 1 0 1 1 1 1
II 1 1 0 0 1 0
III 1 1 1 0 0 0
IV 1 0 0 0 0 1
A) Consider the fuzzy sets 𝐴̃ and 𝐵̃ defined on the interval=[0,5] of real numbers, by
𝑥
the membership grade function µ𝐴̃ (x)= 𝑥+1 µ𝐵̃ (x)=2−𝑥 Determine the mathematical
formulae and graphs of the membership grade functions of each of the following sets :
7. a) 𝐴̃𝑐 b) 𝐵̃𝑐 c) (Ã ∪ B ̃) 𝑐 d) 𝐴̃𝑐 ∪ 𝐵̃ 𝑐 e) A ̃ ∪ B ̃
8 3
OR
B) Write short notes on: i) Partition ii) Covering iii)Rule of addition iv)Rules of
Inclusion and Exclusion
Explain each with examples.

You might also like