SET - A
DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING
(Common to Information Technology)
BATCH (2022 - 2026): III Year/ VI SEM
CCS355 Neural Networks and Deep Learning (R-2021)
MODEL EXAMINATION
DATE: TIME: 3 HRS TOTAL MARKS: 100
PART A: (10X2=20 Marks)
1. What is meant by ANN? Why ANN is used? (C315.1, PO1,PO3,PO5)
2. List the Major application of neural networks (C315.1, PO1,PO3,PO5)
3. What is delta rule for pattern association (C315.2, PO1,PO3)
4. State the advantages and the limitations of Hopfield networks (C315.2, PO1,PO3)
5. What is ELM? (C315.3, PO1,PO3)
6. How pooling handles inputs of varying size. (C315.3, PO1,PO3)
7. What is VC dimension? (C315.4, PO1,PO3)
8. How does regularization helps reduce over fitting? (C315.4, PO1,PO3)
9. Summarize the advantages of recursive nets over recurrent nets. (C315.5, PO1,PO3)
10. Why RNN is suitable for sentiment analysis? (C315.5, PO1,PO3)
PART B: (5X13=65 Marks)
11. a) Describe the basic models of artificial neural networks each with suitable diagram
(C315.1, PO1,PO3,PO5) (13)
(Or)
b) Explain the basic activation functions used in neural networks. (C315.1, PO1,PO3,PO5)
(13)
12. a) What are associative memory networks and what are the two types of associative
memory networks? Explain the training and testing algorithm for hetero associative memory
network.. (C315.2, PO1,PO3) (13)
(Or)
b) What is Hopfield neural networks and what are the types of Hopfield neural networks?
Describe the architecture of discrete Hopfield network and demonstrate training and testing
algorithm of discrete Hopfield network. (C315.2, PO1,PO3) (13)
13. a) what is spiking neural network and how does spiking neural network work? Outline the
advantages and disadvantages of spiking neural network work. (C315.3, PO1,PO3) (13)
(Or)
b) Explain the basic structure of Convolutional Neural Network (CNN) in detail. (C315.3,
PO1,PO3) (13)
14.a) Illustrate back propagation process in deep neural network with chain rule. (C315.4,
PO1,PO3) (13)
(Or)
b) What is regularization for deep learning? Illustrate different regularization techniques in deep
learning. (C315.4, PO1,PO3) (13)
15. a) Illustrate the working of bidirectional recurrent neural network with suitable example.
Also summarize the advantages and disadvantages of recurrent neural network. (C315.5,
PO1,PO3) (13)
(Or)
b) What is an auto encoder and what are the types of autoencoder? Describe the application of
various autoencoders. (C315.5, PO1,PO3) (13)
PART C: (1X15=15 Marks)
16.a) Design a deep learning model using CNN for object detection. Justify usage of size of
kernel filter and dimension of convolution output accordingly. Also clearly explain the model
design. (C315.3, PO1,PO3) (15)
(Or)
b) Design a basic RNN model for sentiment analysis to predict the number of positive negative
reviews based on the sentiments of a movie reviews with appropriate normalization and
optimization techniques. Also clearly explain the model design. (C315.5, PO1,PO3) (15)
PREPARED BY- VALIDATED BY- APPROVED BY-
