Scribd
Upload
173 views3 pages

Shoolini University Mid Sem

This document provides a mid-semester exam for a Deep Learning course, consisting of 20 multiple choice and long answer questions worth a total of 52 marks to be completed within 1 hour and 30 minutes via a Google Form. The questions cover key concepts in deep learning including loss functions, gradient descent, batch normalization, neural network architecture, activation functions, and more.

Uploaded by

Clash Clan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
173 views3 pages

Shoolini University Mid Sem

This document provides a mid-semester exam for a Deep Learning course, consisting of 20 multiple choice and long answer questions worth a total of 52 marks to be completed within 1 hour and 30 minutes via a Google Form. The questions cover key concepts in deep learning including loss functions, gradient descent, batch normalization, neural network architecture, activation functions, and more.

Uploaded by

Clash Clan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Shoolini University Mid Sem-I

Course name: Deep Learning (CSU675)

MCQs - 2 marks each and Long Answer type- 5 marks each. Total : (52 marks)
Time : 1 hour 30 mins
Response Sheet is in the form of google form:
https://docs.google.com/forms/d/e/1FAIpQLSeASeoErzJwpCLenmqBIHgIVvo9oeT5eYu9qXOw
wzU1zcJSOw/viewform?usp=sf_link

1. Why loss function is used?


a. Calculate error value in the forward network
b. Optimize the error values according to the error rate.
c. Both a and b
d. None
2. Which loss function is generally used in regression problems?
a. Cross entropy
b. MSE
c. Logarithmic loss
d. Both b and c
3. Why do we use gradient descent?
a. finds parameter of a model that minimizes the cost
b. Adjust weights at the input layer.
c. Both a and b
d. None
4. Batch normalization helps to prevent
a. Activation function becomes to high or low
b. Training speed to become too slow
c. Both a and b
d. None
5. Suppose the number of nodes in the input layer is 5 and the hidden layer is 10.
The maximum number of connections from the input layer to the hidden layer
would be-
a. 50
b. 25
c. 30
d. None
6. In a neural network, which of the following causes the loss not to decrease
faster?
a. Stucked at local minima
b. Learning rate is too Fast
c. Low regularization rate
d. All
7. Which of the following makes a neural network non-linear?
a. ReLU
b. Batch gradient descent
c. CNN
d. ALL
8. For a binary classification problem, which of the following activation function is
used-
a. Softmax
b. ReLU
c. Sigmoid
A. Either a or c
B. Either B or a
C. Either B and C
D. None
9. Suppose you have a dataset from where you have to predict two classes. Then which
of the following configuration you should use in the output layer?
a. Activation function= softmax, loss function= cross entropy
b. Activation function= sigmoid, loss function= cross entropy
c. Activation function= sigmoid, loss function= MSE
d. None
10. In a classification problem, which of the following activation function is most widely
used in the output layer of neural networks?
a. ReLU
b. Sigmoid
c. Softmax
d. None
11. If you are given a dataset of elliptical shape which visually shows red as class 0
which is the outer circle and green is the class 1 which is the inner circle. Now consider
you are given a neural network that contains only 1 hidden layer and in that layer, you
have only 3 neurons. What would be the nature of the decision boundary?
a. Elliptical
b. Circular
c. Linear
d. Triangle
12. Compute the gradient for the following function: f(x,y)=x3 +yx2 + y2
a. 3x2+2xy, x2+2y
b. 3x2+2x, x2+2y
c. 3x2, 2y
d. 3x+2xy, x+2y
13. Consider the following, X0=3, learning rate=0.01 and f(x)=(x+5)2. What would be the
value of X if it undergoes gradient descent after 2nd iteration?
a. 2.6832
b. 2.6842
c. 2.4832
d. None
14. If you are given 1 neuron with activation function as relu. What would be the output
if the neuron is tuned to w1=0.1,w2=-1,b=14 and given input values as 10, 5.
a. 0
b. 1
c. 99%
d. 88%
15. What is the criteria for a decision boundary to fit ideally on given training samples.
a. Entropy is high
b. Entropy is low
c. Gini index is high
d. Gini is low
16. If you are given 1 neuron with activation function as sigmoid. What would be the
output if the neuron is tuned to w1=0.1,w2=-1,b=14 and given input values as 10, 5.
a. 0
b. 1
c. 99%
d. 88%
17. Why sigmoid is popular? Why not relu?
18. What happens when the training samples are not uniformly distributed for different
classes?
19. Brief the concept of regularization from the topic of regression.
20. Write a short note on cross-entropy.

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy