Skip to content
EndsemMay 20242019 patternTE AI&DS

Deep Learning (DL) SPPU Endsem Question Paper May 2024

All 8 questions from this SPPU endsem paper, as printed. Read them here or open the PDF.

Open the PDF
Subject code
417532D
Pattern
2019 pattern
Exam
Endsem
Session
May 2024
Max. marks
70
Time
2½ Hours
Semester
VIII
Paper code
[6263]-395

Instructions to the candidates

  1. Answer four questions Q1 or Q2, Q3 or Q4, Q5 or Q6, Q7 or Q8.
  2. Neat diagrams must be drawn wherever necessary.
  3. Assume suitable data if necessary.

Questions

  1. Q1

    1. a)

      What is pooling in CNNs? Why is it used? Explain the difference between max pooling and average pooling.

      [9]
    2. b)

      How do you determine the number of filters in each convolutional layer?

      [5]
    3. c)

      Define convolutional neural network (CNN), and how does it differ from other types of neural networks?

      [4]

  2. Q2

    1. a)

      Explain the concept of padding in CNNs. Why is it used? What are strides in CNNs?

      [9]
    2. b)

      Describe the typical architecture of a CNN.

      [5]
    3. c)

      What is the main purpose of using CNNs in deep learning?

      [4]

  3. Q3

    1. a)

      What are some common performance metrics used to evaluate RNNs? How do these metrics differ for different applications of RNNs?

      [9]
    2. b)

      Describe the long short-term memory (LSTM) unit and its components.

      [8]

  4. Q4

    1. a)

      What is the difference between recurrent neural networks (RNNs) and feedforward neural networks?

      [8]
    2. b)

      What are hyperparameters in the context of neural networks? How do researchers select appropriate hyperparameters for training RNNs? B.E. Artificial Intelligence and Data Science

      [9]

  5. Q5

    1. a)

      What is a generative adversarial network (GAN), and how does it work to generate realistic synthetic data?

      [9]
    2. b)

      Describe different types of GANs, How do these types of GANs differ in their architecture and training?

      [9]

  6. Q6

    1. a)

      How are deep generative models used in machine learning and artificial intelligence?

      [8]
    2. b)

      What are some common challenges of using GANs, and how can they be addressed in practice?

      [7]
    3. c)

      What are some applications of GANs in computer vision?

      [3]

  7. Q7

    1. a)

      How can reinforcement learning be applied to play Tic-Tac-Toe? What are the key components of a reinforcement learning algorithm for playing Tic - Tac.

      [8]
    2. b)

      What is deep reinforcement learning, and how does it combine deep learning with reinforcement learning?

      [9]

  8. Q8

    1. a)

      Explain the concept of a Markov Decision Process. What are the main components of an MDP?

      [6]
    2. b)

      Describe the architecture of a deep Q recurrent network.

      [6]
    3. c)

      What are some of the main challenges faced in reinforcement learning? 

      [5]