Roll No.

2olango Exam Code : D-23

Subject Code-16366
B.Tech. (CSE/IT) (Fifth
(Batch 2018 Semester)
onwards)/B.Tech. (CSE)
(Artificial Intelligence and Machine
Learning) (Fifth Semester) (Batch 2021
Onwards) EXAMINATION
MACHINE LEARNING
PCC CSEA-301-T

Time:3Hours Maximum Marks:70

Note : Attempt Five
questions in all, selecting one
question from each Unit. Q. No. 1 is
compulsory. All questions carry equal marks.
1. (i) What is the assumption made about the
distribution of data in each class for
Linear Discriminant Analysis to be
cffective ?
(3-58-4-1023) J-16366
PT.O.
 (i) What is the role of
bias in
learning algorithms ? concept
(ii) Why is

important in dimachine
mensionallearning
ity randeductdataion
analysis ?
(iv) Provide a simple
example of using a
perceptron for binary classification.
(v) Differentiate
between strong and weak
inductive biases.
(vi) Discuss the concepts of the instance space
and version space in the context the
Candidate Elimination algorithm.
(vii) Compare and contrast linear
SVMs with
non-linear SVMs.
Unit I
2. (a)
Define version spaces and their
to
concep learming. How arerelevance
spaces used to version
the represent
learning process ? uncertainty in
J-16366
 (b) Why is the
lack of
data a
concen in diversity in training
Discuss. machine learning ?
3. (a) What is
machine
itdiffer from learning and how docs
Discuss traditional
the broad programmi ng?
scope and
(b)
of
machine learning in various applications
Explain the concepts of domains.
inference in the context training and
learning models. of
machine
Unit II
4. (a)
Compare and contrast the Lcast
Error Thethod with Maximum Squared
in terms of
their approaches to
estWhatimatiison, Liparamet
kelihooder
(b)
a Bayesian Belief Network
(BBN) and how does it represent
probabilistic
variables ? Discuss.
relationships anmong

(3-58-5-1023) J-16366 PLO.

! 3
 What is the Gibbs'
5. (a) algorithm and what is
its primary purpose in statistical and
computational contexts ?
(b) Explain the relationship
between the
Bayes Optimal Classifier and Bayes'
Theorem. How is the classifier influenced
by prior probabilities and
likelihoods ?
Unit III
6. (a) Define the cost function in the
context
of neural networks. How
does Gradient
Descent minimize the cost function
to
optimize neural network
parameters?
(b) What is a
Support Vector Machine
(SVM) ? Explain the concept of a
hyperplane and its role in SVM
classification.

J-16366
7. (a) Explain how Support Vector
can be adapted for Machines
is the
regression tasks. What
loss function used in SYM
regression ?
(b) Compare activation functions sigmoid and
ReLU in terms of their
properties and
use cases.

Unit IV
8. a)
Define the Bellman
the
equation. How does
Bellman equation express the
relationship between Q-values of different
states and actions ? Discuss.
(b) Define eigenvalues and
eigenvectors in
the context of PCA. Explain how eigen
vectors are used to form principal
components.
PT.O.
3-58-6-1023) J-16366
o () Describe the Q-learning algorithm and
its key components.
(b) Explain how a policy and a value function
guide the
decision-making of a

reinforcement learning agent.