X

(https://swayam.gov.in) (https://swayam.gov.in/nc_details/NPTEL)

arjunyamarthy@gmail.com 

NPTEL (https://swayam.gov.in/explorer?ncCode=NPTEL) » Deep Learning - IIT Ropar (course)

If already
registered, click
to check your
Week 6 : Assignment 6
payment status The due date for submitting this assignment has passed.
Due on 2024-03-06, 23:59 IST.

Course Assignment submitted on 2024-03-06, 19:34 IST

outline

About 1) We are using the following autoencoder with linear encoder and linear decoder. The 1 point
NPTEL () eigenvectors associated with the covariance matrix of our data X is (V1 , V2 , V3 , V4 , V5 ) . What
are the representations most likely to be learned by our hidden layer H? (Eigenvectors are
How does an written in decreasing order to the eigenvalues associated with them)
NPTEL
online
course
work? ()

Week 0 ()

Week 1 ()

Week 2 ()

Week 3 ()
V1 , V2

week 4 ()
V4 , V5

Week 5 ()
V1 , V3

Week 6 () V1 , V2 , V3 , V4 , V5

Introduction to
Yes, the answer is correct.
Autoncoders

S 1
(unit? Score: 1
unit=83&lesso Accepted Answers:
n=84) V1 , V2

Link between 2) What is the purpose of a decoder in an autoencoder? 1 point

PCA and
Autoencoders To reconstruct the input data
(unit? To generate new data
unit=83&lesso
To compress the input data
n=85)
To extract features from the input data
Regularization
in
Yes, the answer is correct.
Score: 1
autoencoders
Accepted Answers:
(Motivation)
To reconstruct the input data
(unit?
unit=83&lesso
n=86) 3) What are the advantages of using a denoising autoencoder? 1 point

Denoising Robustness to noisy input data

Autoencoders
Reduction of the risk of overfitting
(unit?
unit=83&lesso Faster training time
n=87) It promotes sparsity in the hidden layer
Sparse No, the answer is incorrect.
Autoencoders Score: 0
(unit? Accepted Answers:
unit=83&lesso Robustness to noisy input data
n=88) Reduction of the risk of overfitting

Contractive
Autoencoders 4) We are given an autoencoder A. The average activation value of neurons in this 1 point
(unit? network is 0.06. The given autoencoder is:
unit=83&lesso
n=89) Contractive autoencoder
Overcomplete neural network
Lecture
Material for Sparse autoencoder
Week 6 (unit? Denoising autoencoder
unit=83&lesso
n=90) Yes, the answer is correct.
Score: 1
Week 6 Accepted Answers:
Feedback Sparse autoencoder
Form: Deep
Learning - IIT 5) Which of the following networks represents an autoencoder? 1 point
Ropar (unit?
unit=83&lesso
n=235)

Quiz: Week 6
: Assignment
6
(assessment?
name=271)

Week 6:
Solution (unit?
unit=83&lesso
n=250)

Week 7 ()

Week 8 ()

Week 9 ()

week 10 ()

Week 11 ()

Week 12 ()

Download
Videos ()

Books ()

Text
Transcripts ()

Problem
Solving
Session -
Jan 2024 ()

No, the answer is incorrect.

S 0
 Score: 0
Accepted Answers:

6) If the dimension of the hidden layer representation is less than the dimension of the 1 point
input layer, then what kind of autoencoder do we have?

Complete autoencoder
Under-complete autoencoder
Overcomplete autoencoder
Sparse autoencoder

Yes, the answer is correct.

Score: 1
Accepted Answers:
Under-complete autoencoder

7) Consider a scenario where we have a single data point with features 1 point
taking values −1, 2, 3.2, −5.6, 0 respectively. In this case, which function
x1 , x2 , x3 , x4 , x5

would be most suitable to use on the output layer (decoder) of an autoencoder for training on this
particular dataset?

Logistic
Relu
Tanh
Linear

Yes, the answer is correct.

Score: 1
Accepted Answers:
Linear
 8) If the dimension of the input layer in an over-complete autoencoder is 5, what is the 1 point
possible dimension of the hidden layer?

4
2
8
0

Yes, the answer is correct.

Score: 1
Accepted Answers:
8

9) What is the primary objective of contractive autoencoders that distinguishes them 1 point
from vanilla autoencoders?

They learn a low-dimensional representation of the input data

They minimize the reconstruction error between the input and the output
They capture only the important variations/features in the data
They maximize the mutual information between the input and the output

Yes, the answer is correct.

Score: 1
Accepted Answers:
They capture only the important variations/features in the data

10) What are the possible applications of autoencoders? 0 points

Data Compression
Extraction of important features
Reducing noise
All of these

Yes, the answer is correct.

Score: 0
Accepted Answers:
All of these