DR.
BABASAHEB AMBEDKAR TECHNOLOGICAL UNIVERSITY, LONERE
Regular\Supplementary Winter Examination – 2024
Course: B.Tech Branch : Computer Engineering/Computer Science & Engg
Semester : VII
Subject Code & Name: BTCOE705B , Deep Learning
Max Marks: 60 Date: 21/02/2025 Duration: 3 Hr.
51647754
51647754
51647754
Instructions to the Students:
1. Each question carries 12 marks.
2. Question No. 1 will be compulsory and include objective-type questions.
3. Candidates are required to attempt any four questions from Question No. 2 to Question No. 6.
4. The level of question/expected answer as per OBE or the Course Outcome (CO) on which the
question is based is mentioned in ( ) in front of the question.
5. Use of non-programmable scientific calculators is allowed.
6. Assume suitable data wherever necessary and mention it clearly.
(Level/CO) Marks
Q. 1 Objective type questions. (Compulsory Question) 12
1 What is the purpose of an activation function in a neural network? Understand 1
a. introduces b. determines the c. helps in back d. All of these
non-linearity to output of a propagation
51647754
51647754
51647754
the network neuron
2 What is the role of gradient descent in deep learning? Analyze 1
a. To minimize b. To find the c. To update the d. All of these
the loss optimal weights network's
function for the neural parameters
network
3 What is the purpose of dropout regularization in deep learning? Analyze 1
a. To reduce b. To increase the c. To improve d. To handle
overfitting model's capacity the training speed imbalanced
datasets
4 Which activation function is commonly used in the hidden layers of a deep Design 1
51647754
51647754
51647754
neural network?
a. Rectified b. Sigmoid c. Tanh d. Softmax
Linear Unit
5 Which of the following is a common optimization algorithm used in deep Understand 1
learning?
a. Gradient b. Stochastic c. Adam d. All of these
Descent Gradient Descent
6 What is the purpose of a recurrent neural network (RNN)? Understand 1
a. To handle b. To classify c. To perform d. To generate
sequential data images dimensionality synthetic data
reduction
7 Which deep learning technique is used for generating new, realistic data Understand 1
samples?
a. Generative b. Convolutional c. Reinforcement d. Transfer
Adversarial Neural Networks Learning Learning
Networks
8 Which deep learning technique is used for unsupervised feature learning? Analyze 1
a.Autoencoders b. Convolutional c. Recurrent d.
51647754
51647754
51647754
Neural Networks Neural Networks Reinforcement
(CNNs) (RNNs) Learning
9 Which of the following is a common metric used to evaluate the Understand 1
performance of a classification model in deep learning?
a. Accuracy b. Mean Absolute c. Mean Squared d. R-squared
Error (MAE) Error (MSE)
10 CNN is mostly used when there is ? Understand 1
a. structured b. unstructured c. Both A and B d. None of
data data these
11 Which of the following is/are Limitations of deep learning? Understand 1
51647754
51647754
51647754
a. Data labeling b. Obtain huge c. Both A and B d. None of the
training datasets above
12 The number of nodes in the input layer is 10 and the hidden layer is 5. The Understand 1
maximum number of connections from the input layer to the hidden layer
are
a. 50 b. less than 50 c. more than 50 d. It is an
arbitrary value
Q. 2 Solve the following. 12
A) Explain the working of deep learning model. Understand 6
51647754
51647754
51647754
B) Distinguish between Deep Learning and Machine Learning. Analyze 6
Q.3 Solve the following. 12
A) What is neuron? Explain structure of biological neuron. Remember 6
B) Explain architecture of feed forward neural network with necessary Understand 6
convention.
Q. 4 Solve Any Two of the following. 12
A) What is back propagation learning ? Remember 6
B) How weights are updated at output layer for multilayer neural network ? Apply 6
C) Explain the concept of layer by layer pretraining mechanism. Understand 6
12
Q.5 Solve Any Two of the following.
A) Explain different variants of auto encoder. Understand 6
B) Explain architecture of Convolutional Neural Network . Remember 6
C) What are the advantages of Convolutional Neural Network over multilayer Understand 6
perceptron?
51647754
51647754
51647754
Q. 6 Solve Any Two of the following. 12
A) Explain recurrent neural network. Understand 6
B) What is overfitting and underfitting problem?How is it resolved in neural Remember 6
networks?
C) What is the need of normalization? Explain different kinds of Understand 6
normalization techniques in neural networks.
*** End ***
51647754
51647754
51647754
51647754
51647754
51647754
