Consider a simple neural network with a hidden unit with ReLU as an activation function and a linear output unit. The loss used is the squared loss: Figure from the original question paper . Figure from the original question paper Question text from the original paper, with its maths as pictures

(B) Question text from the original paper, with its maths as pictures

(C) Question text from the original paper, with its maths as pictures

(D) Question text from the original paper, with its maths as pictures

(F) Question text from the original paper, with its maths as pictures

(A) Question text from the original paper, with its maths as pictures

(E) Question text from the original paper, with its maths as pictures

Consider a simple neural network with a hidden unit with ReLU as an activation function and a linear output unit. The loss used is the squared loss: Figure from the original question paper . Figure from the original question paper Question text from the original paper, with its maths as pictures

(B) Question text from the original paper, with its maths as pictures

(C) Question text from the original paper, with its maths as pictures

(D) Question text from the original paper, with its maths as pictures

(F) Question text from the original paper, with its maths as pictures

(A) Question text from the original paper, with its maths as pictures

(E) Question text from the original paper, with its maths as pictures

Consider a neural network with one hidden layer as shown below: Figure from the passage in the original paper Question text from the original paper, with its maths as pictures Figure from the passage in the original paper . Based on the above data, answer the given subquestions. Figure from the passage in the original paper Figure from the passage in the original paper Based on the previous questions 11 and 12, select the correct options from the following:

(A) Question text from the original paper, with its maths as pictures

(C) Question text from the original paper, with its maths as pictures

(B) Gradient updates will be larger in case of regularised loss.

(D) Question text from the original paper, with its maths as pictures

Consider a neural network with one hidden layer as shown below: Figure from the passage in the original paper Question text from the original paper, with its maths as pictures Figure from the passage in the original paper . Based on the above data, answer the given subquestions. Figure from the passage in the original paper Figure from the passage in the original paper Based on the previous questions 11 and 12, select the correct options from the following:

(A) Question text from the original paper, with its maths as pictures

(C) Question text from the original paper, with its maths as pictures

(B) Gradient updates will be larger in case of regularised loss.

(D) Question text from the original paper, with its maths as pictures

Question text from the original paper, with its maths as pictures Figure from the passage in the original paper Based on the above data, answer the given subquestions. Question text from the original paper, with its maths as pictures

(B) Question text from the original paper, with its maths as pictures

(A) It increases the spatial resolution of the feature map.

(C) It allows the network to learn parallel features at different scales.

(D) It is only used to add non-linearity and does not affect the parameter count.

Question text from the original paper, with its maths as pictures Figure from the passage in the original paper Based on the above data, answer the given subquestions. Question text from the original paper, with its maths as pictures

(B) Question text from the original paper, with its maths as pictures

(A) It increases the spatial resolution of the feature map.

(C) It allows the network to learn parallel features at different scales.

(D) It is only used to add non-linearity and does not affect the parameter count.

January 2026 term · Deep Learning · BSCS3004

Deep Learning Quiz 2: 12 April 2026 (January 2026 term)

The IIT Madras BS Deep Learning (Deep Learning) Quiz 2 paper sat on 12 Apr 2026, in the January 2026 term: 30 questions for 80 marks in 120 minutes. Every question is below with its answer. Take it as a timed mock test to be marked, or read it through first.

Questions: 30
Marks: 80
Duration: 120 min
MCQ: 8
MSQ: 6
Numerical: 16

Take as mock test Practise with answers

Updated 25 Sept 2026

Official paper: Deep Learning 07 Apr 26 · No negative marking.

Question 1

+1 markOne correct option

Which of the following optimization algorithms combines Nesterov accelerated gradient with adaptive moment estimation?

A
NAdam
B
Adam
C
RMSprop with momentum
D
Nesterov accelerated gradient descent

Show answer

Correct answer

A
NAdam

Question 1 on its own page

Question 2

+1 markOne correct option

Which of the following optimization algorithms combines Nesterov accelerated gradient with adaptive moment estimation?

A
NAdam
B
Adam
C
RMSprop with momentum
D
Nesterov accelerated gradient descent

Show answer

Correct answer

A
NAdam

Question 2 on its own page

Question 3

+2 marksOne correct option

Suppose that we apply Dropout regularization to a feed forward neural network. Suppose further that mini-batch gradient descent algorithm is used for updating the parameters of the network. Choose the correct statement from the following:

A
This will lead to sparcity in the trained weights.
B
C
D
None of these.

Show answer

Correct answer

Question 3 on its own page

Question 4

+2 marksOne correct option

A
This will lead to sparcity in the trained weights.
B
C
D
None of these.

Show answer

Correct answer

Question 4 on its own page

Question 5

+4 marksOne correct option

Match the optimization algorithms in Column A with their corresponding properties or characteristics in Column B. Note that an algorithm in Column A may correspond to multiple properties in Column B.

Choose the correct set of mappings:

A
1-(i), 2-(v), 3-(ii, iii), 4-(iv)
B
1-(ii), 2-(iv), 3-(iii), 4-(i, v)
C
1-(ii), 2-(iii), 3-(i, iv), 4-(v)
D
1-(v), 2-(i), 3-(iv), 4-(ii, iii)

Show answer

Correct answer

C
1-(ii), 2-(iii), 3-(i, iv), 4-(v)

Question 5 on its own page

Question 6

+4 marksOne correct option

Match the optimization algorithms in Column A with their corresponding properties or characteristics in Column B. Note that an algorithm in Column A may correspond to multiple properties in Column B.

Choose the correct set of mappings:

A
1-(i), 2-(v), 3-(ii, iii), 4-(iv)
B
1-(ii), 2-(iv), 3-(iii), 4-(i, v)
C
1-(ii), 2-(iii), 3-(i, iv), 4-(v)
D
1-(v), 2-(i), 3-(iv), 4-(ii, iii)

Show answer

Correct answer

C
1-(ii), 2-(iii), 3-(i, iv), 4-(v)

Question 6 on its own page

Question 7

+2 marksOne or more correct options

Consider training a deep neural network for a multi-class classification problem. The output layer uses softmax activation, and the hidden layers use ReLU as an activation function. After a few training iterations, you observe that the weights in the initial hidden layers barely get updated. Which of the following actions could help tackle this issue?

Select all that apply.

A
Replace ReLU with Leaky ReLU in the hidden layers.
B
Replace ReLU by tanh in the hidden layers.
C
Increasing the number of neurons in the output layer.
D
Increasing the learning rate.

Show answer

Correct answers

A
Replace ReLU with Leaky ReLU in the hidden layers.
D
Increasing the learning rate.

Question 7 on its own page

Question 8

+2 marksOne or more correct options

Select all that apply.

A
Replace ReLU with Leaky ReLU in the hidden layers.
B
Replace ReLU by tanh in the hidden layers.
C
Increasing the number of neurons in the output layer.
D
Increasing the learning rate.

Show answer

Correct answers

A
Replace ReLU with Leaky ReLU in the hidden layers.
D
Increasing the learning rate.

Question 8 on its own page

Question 9

+4 marksOne or more correct options

Consider a simple neural network with a hidden unit with ReLU as an activation function and a linear output unit. The loss used is the squared loss:

Select all that apply.

Show answer

Correct answers

Question 9 on its own page

Question 10

+4 marksOne or more correct options

Consider a simple neural network with a hidden unit with ReLU as an activation function and a linear output unit. The loss used is the squared loss:

Select all that apply.

Show answer

Correct answers

Question 10 on its own page

Question 11

+3 marksNumerical answer

Show answer

Correct answer: 0

Question 11 on its own page

Question 12

+3 marksNumerical answer

Show answer

Correct answer: 0

Question 12 on its own page

Question 13

+3 marksNumerical answer

Show answer

Correct answer: 0.4

Question 13 on its own page

+4 marksNumerical answer

Consider a neural network with one hidden layer as shown below:

. Based on the above data, answer the given subquestions.

Show answer

Correct answer: 0.35 (accepted within ±0.02)

Question 19 on its own page

+3 marksOne or more correct options

Consider a neural network with one hidden layer as shown below:

. Based on the above data, answer the given subquestions.

Based on the previous questions 11 and 12, select the correct options from the following:

Select all that apply.

A
B
Gradient updates will be larger in case of regularised loss.
C
D

Show answer

Correct answers

Question 23 on its own page

Question 24

+3 marksOne or more correct options

Consider a neural network with one hidden layer as shown below:

. Based on the above data, answer the given subquestions.

Based on the previous questions 11 and 12, select the correct options from the following:

Select all that apply.

A
B
Gradient updates will be larger in case of regularised loss.
C
D

Show answer

Correct answers

Question 24 on its own page

Question 25

+2 marksNumerical answer

Based on the above data, answer the given subquestions.

Calculate the total number of learnable parameters in Module A.

Show answer

Correct answer: 40960

Question 25 on its own page

Question 26

+2 marksNumerical answer

Based on the above data, answer the given subquestions.

Calculate the total number of learnable parameters in Module A.

Show answer

Correct answer: 40960

Question 26 on its own page

Question 27

+2 marksNumerical answer

Based on the above data, answer the given subquestions.

Calculate the total number of learnable parameters in Module B.

Show answer

Correct answer: 151552

Question 27 on its own page

Question 28

+2 marksNumerical answer

Based on the above data, answer the given subquestions.

Calculate the total number of learnable parameters in Module B.

Show answer

Correct answer: 151552

Question 28 on its own page

Question 29

+3 marksOne correct option

Based on the above data, answer the given subquestions.

A
It increases the spatial resolution of the feature map.
B
C
It allows the network to learn parallel features at different scales.
D
It is only used to add non-linearity and does not affect the parameter count.

Show answer

Correct answer

Question 29 on its own page

Question 30

+3 marksOne correct option

Based on the above data, answer the given subquestions.

A
It increases the spatial resolution of the feature map.
B
C
It allows the network to learn parallel features at different scales.
D
It is only used to add non-linearity and does not affect the parameter count.

Show answer

Correct answer

Question 30 on its own page