ANN AND FUZZY LOGIC
SEM-VII, 2012-13
B.TECH EXAMINATION
UTTARAKHAND TECHNICAL UNIVERSITY
(UTU)
Time: 3 hours
Total
Marks: 100
Note: Attempt all questions.
Answer any four:
- What is a biological neuron? Given its structure and
explain its functioning.
- Explain McCulloh Pitts model of artificial neuron.
- Using artificial neuron model with 'Threshold' function
show its use as 'OR' and 'AND' gates.
- Differentiate between a single layer perception and a
multi layer perception.
- Discuss log sigmoid and tan sigmoid activation
functions.
- Draw an MLP of [5, 3, 2] structure and mark the
directions of signal and error flow.
Answer any four:
- Draw any four types of membership functions and give
their mathematical expressions.
- How is a fuzzy set represented? Define support, height
and compliment of a fuzzy set.
- If A = [a1, a2] and B = [b1,
b2] are two fuzzy sets, what will be the resulting sets after
performing addition, subtraction, multiplication and division operations
on the sets?
- Describe briefly Mamdani's inference system.
- What do you understand by 'Cartesian product' of two
fuzzy sets? If A = {0.2/x1, 0.5/x2, 1/x3}
and B = {0.3/y1, 0.9/y2} find the Cartesian product
between A and B.
- What is Defuzzification? Give various methods of
defuzzification.
Answer any two:
- What is error back propagation? In a MLP, how is
learning achieved at output and hidden layer? Derive the expression for
the weight correction.
- What are various factors that affect learning in an
error back propagation net? Discuss effect of each of them.
- Why an XOR problem needs a multilayer perceptron to
solve it? Design a net using threshold functions to describe an XOR gate.
Answer any two:
- Through a block diagram explain various components of a
fuzzy logic controller. Name various applications where it is used.
- What is fuzzy 'if then rules'? How do these differ in
Mamdani's FIS and T.Sugeno FIS?
- What is a neuro-fuzzy interference system? Discuss its
advantages over the individual systems.
Answer any two:
- Illustrate briefly how ANN can be used for an image
processing.
- In the ANN shown, calculate outputs of neurons
(i) forward pass
(ii) reverse pass, given activation function as sigmoid function and target output of 0.5.
- Fuzzy theory has generally been confused with the
probability theory differentiate between the two.
_____________utu previous year question papers of B.tech, Bba, B.com, B.sc and Mba old year question papers Uttarakhand technical University (UTU)____________
No comments:
Post a Comment