QUESTION BANK
Name of the Department: COMPUTER SCIENCE AND ENGINEERING
Name of the Subject: DEEP LEARNING
Subject Code: 23CS2902
Semester: 03
Unit I – BASICS OF NEURAL NETWORKS
PART-A (2 Marks)
[Link]. Question Std. of
question
Cognitive
Level
1. List The Types of Deep Learning Networks? LO
2. What are Deep Belief Networks and write its steps to perform DBN? LO
3. What are autoencoders? LO
4. What is a Neuron in Deep Learning? IO
5. What is an Activation Function and its 6 types? IO
6. What is the Perceptron model in ML? IO
7. What is a binary classifier in ML? IO
8. List the basic components of perceptron? IO
9. How does perceptron work? IO
10. List the types of perceptron model and explain it? IO
Part – B & C (13/ 15/ 16 marks)
[Link]. Question Std. of Mark
question
Cognitive
Level
1. Explain the basic architecture of neurons? LO 13
2. Explain perceptron algorithm? LO 15
3. Explain the feed forward and back propagation networks? IO 13
Unit II – INTRODUCTION TO DEEP LEARNING
PART-A (2 Marks)
[Link]. Question Std. of
question
Cognitive
Level
1. What is gradient descent? LO
2. Why need back propagation algorithm? LO
3. How to resolve the vanishing gradient problem? LO
4. What are the types of bias? IO
5. What are the mitigation techniques? IO
6. What is Relu, and what are its advantages? IO
Part – B (13/15/16 marks
[Link]. Question Std. of Mark
question
Cognitive
Level
1. What is RElu and dying RElu? Explain the causes of dying RElu and discuss LO 13
how to rectify it?
2. Explain regularization techniques. LO 13
Unit III – CONVOLUTIONAL NEURAL NETWORKS
PART-A (2 Marks)
[Link]. Question Std. of question
Cognitive
Level
1. What are the applications of CNN? IO
2. What are pooling layers? IO
3. Why are pooling layers needed? IO
4. Explain transfer learning algorithm? IO
Part – B (13/15/16 marks)
[Link] Question Std. of Mark
question
Cognitive
Level
1. Explain the architecture of CNN LO 13
2. Explain image classification using Transfer learning algorithm. LO 13
3. What is transfer learning algorithm? Explain its working procedure and discuss how IO 13
to use it.
Unit IV - MORE DEEP LEARNING ARCHITECTURES
PART-A (2 Marks)
S.N Question Std. of
o question
. Cognitive
Level
1. The architecture diagram for denoising. LO
2. List the applications of LSTM LO
3. What are Bidirectional LSTM LO
4. What is a denoising encoder? IO
5. Define GRU. IO
6. What is contractive Autoencoder? IO
7. Difference between Contractive and variational Autoencoders. IO
8. Define AGN IO
Part – B (13/15/16 marks)
[Link] Question Std. of Mark
. question
Cognitive
Level
1. Illustrate about Encoder, Decoder architectures in detail. LO 13
2. Explain about LSTM architecture in detail. LO 13
3. Describe in detail about the LSTM applications. IO 13
4. Define auto-encoder and its types in Detail. IO 15
5. Explain working of GRU in detail. HO 13
Unit V - APPLICATIONS OF DEEP LEARNING
PART-A (2 Marks)
[Link]. Question Std. of question
Cognitive
Level
1. What is image segmentation? LO
2. What is Automatic image captioning? LO
3. Define object detection. LO
4. What is the LSTM model, and how does it differ from conventional LSTM? IO
5. Describe generative adversarial networks.
6. State about opinion mining. IO
7. Difference between RNN and CNN.
8. What is RNN?
9. What is Tokenization?
10. What is Name and Entity?
Part – B (13/15/16 marks)
[Link] Question Std. of Mark
. question
Cognitive
Level
1. Describe image segmentation in detail. LO 13
2. Describe object detection and explain it in detail. LO 13
3. Explain about Parsing and Sentiment analysis in detail. IO 13
4. Explain in detail about Named Entity Recognition. IO 13
5. Describe in detail Opinion Mining using a Recurrent Neural Network. HO 13
6. Describe Sentence Classification using CNN. HO 13
7. Explain about Dialogue generation with LSTM. 13
8. Describe in detail using Attention models for Computer Vision. 13
9. Explain Automatic Image Captioning. 13
10. Describe Image generation with Generative Adversarial Networks. 13
11. Imagine the task of determining whether a product’s review is positive 13
or [Link] could do it yourself just by reading it, right? But what
happens when the company you work for sells 2k products every single
day? Are you pretending to read all the reviews and manually classify
them? Let’s be honest, your job would be the worst ever. That’s where
Sentiment Analysis comes in and makes your life and job easier.
PART -C
1 Case Study:
[Link] Entity Recognition
[Link] Mining using Recurrent Neural Networks
[Link] and Sentiment Analysis using Recursive Neural Networks
[Link] Classification using Convolutional Neural Networks