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 3 : Assignment
payment status The due date for submitting this assignment has passed.
Due on 2024-02-14, 23:59 IST.

Course Assignment submitted on 2024-02-14, 21:14 IST

outline
Use the following data to answer the questions below The following diagram represents a neural
network containing two hidden layers and one output layer. The input to the network is a column
About vector x ∈ R .
3
The activation function used in hidden layers is sigmoid. The output layer doesn’t
NPTEL ()
contain any activation function and the loss used is squared error loss (predy 2
− truey ) .

How does an
NPTEL
online
course
work? ()

Week 0 ()

Week 1 ()

Week 2 () The following network doesn’t contain any biases and the weights of the network are given
below:
Week 3 ()

⎡ 1 1 2 ⎤
Feedforward 1 1 2
Neural W1=⎢ 3 1 1 ⎥ W2=[ ] W3=[ 2 5 ]
⎣ ⎦ 3 1 1
Networks 1 2 3

(a.k.a ⎡ 1 ⎤
multilayered The input to the network is: x=⎢ 1 ⎥
network of ⎣ ⎦
1
neurons) (unit?
unit=46&lesso
n=47)
Learning The target value y is: y=10
Paramters of
Feedforward
1) What is the total number of parameters in the following network? 1 point
Neural
Networks
15
(Intuition)
(unit? 7
unit=46&lesso 9
n=48)
17
Output
Yes, the answer is correct.
functions and Score: 1
Loss functions
Accepted Answers:
(unit?
17
unit=46&lesso
n=49)
2) What is the predicted output for the given input x1 after doing the forward pass? 1 point
Backpropagati (Choose the option closest to your answer)
on (Intuition)
(unit? 7.33
unit=46&lesso
6.92
n=50)
6.31
Backpropagati
8
on: Computing
Gradients w.r.t. Yes, the answer is correct.
the Output Score: 1
Units (unit? Accepted Answers:
unit=46&lesso 6.92
n=51)
3) Compute and enter the loss between the output generated by input x and the true output y.
Backpropagati
(NAT)
on: Computing
Gradients w.r.t.
9.44
Hidden Units
(unit? Yes, the answer is correct.
unit=46&lesso Score: 1
n=52) Accepted Answers:
(Type: Range) 9.38,9.58
Backpropagati
on: Computing 1 point
Gradients w.r.t.
Parameters 4) If we call the predicted y as y
^ then what is the gradient dL/dy
^ ? (L is the loss 1 point
(unit? function)
unit=46&lesso
n=53) -5.17
-7.52
Backpropagati
on: Pseudo -6.15
code (unit? -7.15
unit=46&lesso
n=54) Yes, the answer is correct.
Score: 1
Derivative of Accepted Answers:
the activation -6.15
function (unit?
unit=46&lesso
5) What is the sum of elements of ∇ω3? (Choose the closest value to your answer) 1 point
n=55)
-12.9
Information -11.6
content,
-10.07
Entropy &
cross entropy -12.14
(unit? Yes, the answer is correct.
unit=46&lesso Score: 1
n=56) Accepted Answers:
-12.14
Lecture
Material for 6) What is the sum of elements of ∇ω2?
Week 3 (unit?
unit=46&lesso -36.15
n=57)
No, the answer is incorrect.
Week 3 Score: 0
Feedback Accepted Answers:
Form: Deep (Type: Range) -1.4,-1.2
Learning - IIT
Ropar (unit? 0 points
unit=46&lesso
7) What is the sum of elements of ∇ω2?
n=186)

Week 3:
-108.45
Solution (unit?
No, the answer is incorrect.
unit=46&lesso Score: 0
n=247) Accepted Answers:
Quiz: Week 3
(Type: Range) -0.08,-0.04
: Assignment 0 points
(assessment?
name=265) 8) The probability of all the events x 1 , x 2 , x 2 . . . . x n in a system is equal (n > 1) . 1 point
What can you say about the entropy H (X) of that system?(base of log is 2)
week 4 ()

H (X) ≤ 1
Week 5 ()

H (X) = 1
Week 6 ()
H (X) ≥ 1
Week 7 ()
We can’t say anything conclusive with the provided information

Week 8 () No, the answer is incorrect.

Score: 0
Accepted Answers:
Week 9 () H (X) ≥ 1

week 10 () 9) Let p and q be two probability distributions. Under what conditions will the cross 1 point
entropy between p and q be minimized?
Week 11 ()

p = q
Week 12 ()
All the values in p are lower than corresponding values in q
Download
Videos () All the values in p are lower than corresponding values in q

Books () p = 0 [0 is a vector]

Y h i
 Yes, the answer is correct.
Text Score: 1
Transcripts () Accepted Answers:
p = q

Problem 10) Suppose we have a problem where data x and label y are related by y = x 2 + 1. 1 point
Solving Which of thefollowing is not a good choice for the activation function in the hidden layer if the
Session - activation function at the output layer is linear?
Jan 2024 ()
Linear
Relu
Sigmoid

−1
Tan (x)

Yes, the answer is correct.

Score: 1
Accepted Answers:
Linear