10/24/25, 9:54 AM Machine Learning and Deep Learning - Fundamentals and Applications - - Unit 7 - Week 4: Perceptron Criteria and
ria and Discrimin…
(https://swayam.gov.in) (https://swayam.gov.in/nc_details/NPTEL)
[email protected]
NPTEL (https://swayam.gov.in/explorer?ncCode=NPTEL) » Machine Learning and Deep Learning -
Fundamentals and Applications (course)
Course Assignment 4
outline The due date for submitting this assignment has passed.
Due on 2025-08-20, 23:59 IST.
About
NPTEL ()
Assignment submitted on 2025-08-19, 12:29 IST
How does an 1) For a dataset that is not linearly separable, what is the expected behavior of the 1 point
NPTEL perceptron learning algorithm?
online
course It will converge to a solution that minimizes the classification error
work? () It will oscillate indefinitely without converging
It will automatically switch to a nonlinear decision boundary
Week 0:
Prerequisite It will reduce the learning rate until convergence is achieved
() Yes, the answer is correct.
Score: 1
Week 1: Accepted Answers:
Introduction It will oscillate indefinitely without converging
()
2) Which of the following is a key difference between the Perceptron Criterion and the 1 point
Week 2: SVM optimization objective?
Bayesian
Decision Perceptron minimizes the margin, while SVM maximizes it
Theory () Perceptron uses a soft margin, while SVM uses a hard margin
Perceptron penalizes all misclassified points equally, while SVM uses slack variables
Week 3-
Perceptron requires a kernel function, while SVM does not
Parametric
Approaches Yes, the answer is correct.
Score: 1
and Non-
Parametric Accepted Answers:
Perceptron penalizes all misclassified points equally, while SVM uses slack variables
https://onlinecourses.nptel.ac.in/noc25_ee181/unit?unit=45&assessment=195 1/4
10/24/25, 9:54 AM Machine Learning and Deep Learning - Fundamentals and Applications - - Unit 7 - Week 4: Perceptron Criteria and Discrimin…
Density 3) What is the effect of using a Radial Basis Function (RBF) kernel in an SVM when 1 point
Estimation () the data is linearly separable in the original space?
Week 4: It reduces the number of support vectors
Perceptron It ensures the decision boundary remains linear
Criteria and
It minimizes the hinge loss more effectively
Discriminativ
e Models () It may overfit the data by creating a more complex decision boundary
Yes, the answer is correct.
Lec 16: Linear Score: 1
Discriminant Accepted Answers:
Functions and It may overfit the data by creating a more complex decision boundary
Perceptron
Criteria (Part I)
4) Which of the following statements is true about the dual form of the SVM 1 point
(unit?
optimization problem?
unit=45&lesso
n=46)
It depends only on the inner products of training samples
Lec 17: Linear It requires the data to be linearly separable in the original space
Discriminant
It eliminates the need for support vectors
Functions and
Perceptron
It directly computes the weight vector w in the primal space
Criteria (Part
II) (unit? Yes, the answer is correct.
unit=45&lesso Score: 1
n=47) Accepted Answers:
It depends only on the inner products of training samples
Lec 18: Linear
Discriminant
Functions and 5) In a one-vs-one (OvO) multi-class classification setup with K classes, how many 1 point
Perceptron binary classifiers (or decision boundaries) are required?
Criteria (Part
III) (unit?
K
unit=45&lesso
n=48) K − 1
Lec 19:
K(K−1)
Support Vector
2
Machine (Part
I) (unit? 2K
unit=45&lesso
n=49)
Yes, the answer is correct.
Score: 1
Lec 20: Accepted Answers:
Support Vector K(K−1)
2
Machine (Part
II) (unit?
6) In multi-class classification, what is a key advantage of using multiple linear 1 point
unit=45&lesso
n=50) discriminant functions gi (x) one for each class?
Lecture Notes It reduces the number of parameters needed.
(unit?
It allows classification even when classes are not linearly separable.
unit=45&lesso
n=91) It ensures that all decision boundaries are nonlinear.
Feedback form It avoids ambiguity by assigning each input to the class with the maximum gi (x) .
(unit?
Yes, the answer is correct.
https://onlinecourses.nptel.ac.in/noc25_ee181/unit?unit=45&assessment=195 2/4
10/24/25, 9:54 AM Machine Learning and Deep Learning - Fundamentals and Applications - - Unit 7 - Week 4: Perceptron Criteria and Discrimin…
unit=45&lesso Score: 1
n=190) Accepted Answers:
It avoids ambiguity by assigning each input to the class with the maximum gi (x) .
Quiz:
Assignment 4 7) Let wij represents weight between node i at layer k and node j at layer (k-1) of a 1 point
(assessment? given multilayer perceptron. The weight updation using gradient descent method is given by: (
name=195) α and E represent learning rate and Error in the output respectively)
Assignment
Solution (unit? ∂E
Wij (t + 1) = Wij (t) + α , 0 ≤ α ≤ 1
unit=45&lesso ∂Wij
n=192)
∂E
Wij (t + 1) = −α , 0 ≤ α ≤ 1
∂Wij
Week 5:
Logistic ∂E
Wij (t + 1) = Wij (t) − α , 0 ≤ α ≤ 1
Regression, ∂Wij
Decision
∂E
Trees and Wij (t + 1) = α
∂Wij
, 0 ≤ α ≤ 1
Hidden
Yes, the answer is correct.
Markov Score: 1
Model ()
Accepted Answers:
∂E
Wij (t + 1) = Wij (t) − α , 0 ≤ α ≤ 1
Week 6 - ∂Wij
Ensemble
Methods () 8) A 3-input neuron has weights 2, 5, and 7. The inputs to the neuron are 4, 6, and 8 1 point
respectively. The transfer function is linear with a constant of proportionality equal to 2. What
Week 7: will be the output?
Dimensionali
188
ty Problem ()
240
Week 8: 116
Mixture 152
Model and
Clustering () Yes, the answer is correct.
Score: 1
Accepted Answers:
Week 9:
188
Clustering ()
9) A set of training samples are given below- 0 points
Week 10:
Artificial
Neural
Network ()
Week 11:
Introduction
to Deep
Neural
Networks () Using Support vector machine algorithm, the Marginal line for the classification can be calculated
as-
Week 12:
Recent
−6.67x 1 + 8.134x 2 − 9.09 = 0
Trends in
Deep 8.21x 1 + 7.12x 2 − 9.09 = 0
Learning ()
https://onlinecourses.nptel.ac.in/noc25_ee181/unit?unit=45&assessment=195 3/4
10/24/25, 9:54 AM Machine Learning and Deep Learning - Fundamentals and Applications - - Unit 7 - Week 4: Perceptron Criteria and Discrimin…
−5.32x 1 − 7.193x 2 + 9.09 = 0
Live session
() −7.21x 1 − 9.173x 2 + 9.09 = 0
No, the answer is incorrect.
Transcript () Score: 0
Accepted Answers:
Books () −7.21x 1 − 9.173x 2 + 9.09 = 0
10) In refer to Q.9, A new test sample (0.5,0.5) is found. The class of the given sample 0 points
is-
Negative
Positive
Both class
Can't say
No, the answer is incorrect.
Score: 0
Accepted Answers:
Positive
https://onlinecourses.nptel.ac.in/noc25_ee181/unit?unit=45&assessment=195 4/4
