Skip to content
EndsemMay 20242019 patternBE E&TC

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
404185C
Pattern
2019 pattern
Exam
Endsem
Session
May 2024
Max. marks
70
Time
2½ Hours
Semester
VII
Paper code
[6263]-134

Printed for: B.E. (E & TC) (Endsem)

Instructions to the candidates

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

Questions

  1. Q1

    1. a)

      What is the significance of Batch Normalization? How covariate shift reduced using batch normalization?

      [7]
    2. b)

      Describe the following activation functions : i) Sigmoid, ii) Threshold, iii) ReLU ‘Vanishing Gradient’ occur in Sigmoid Function and ‘Dying ReLU’ occur in ReLu - Justify

      [10]

  2. Q2

    1. a)

      Write a Note on : i) Dropout Method and ii) Regularization

      [7]
    2. b)

      Explain the concept of Overfitting and Underfitting in deep learning. Support both terms using real time examples?

      [10]

  3. Q3

    1. a)

      With the help of architectures show that Speed and Accuracy of VGG is greater than AlexNet.

      Also asked:Nov 2023

      [10]
    2. b)

      Show that CNN works same as human brain for image recognition through various steps.

      [8]

  4. Q4

    1. a)

      With the help of architecture explain each block of DenseNet in detail. Enlist advantages and disadvantages.

      Also asked:Nov 2023

      [10]
    2. b)

      What is weight initialization? Describe the various weight initialization techniques.

      [8]

  5. Q5

    1. a)

      Explain the working of RNN with suitable diagram. Illustrate the Vanishing Gradient problem occurs in simple RNN.

      Also asked:Nov 2023

      [10]
    2. b)

      How Name Entity Recognition Problem is fixed using Bidirectional RNN? Explain with the help of suitable diagram.

      [7]

  6. Q6

    1. a)

      How short term memory problem avoided in LSTM? Explain with the help of suitable diagram.

      [10]
    2. b)

      What is the difference between LSTM and GRU? Explain the working of GRU with suitable diagram.

      [7]

  7. Q7

    1. a)

      What is NLP? Enlist the advantages and disadvantages of NLP.

      [6]
    2. b)

      What is sentiment analysis? Describe the various use cases of sentiment analysis.

      [6]
    3. c)

      How NLP works in text pre-processing.

      [6]

  8. Q8

    1. a)

      Differentiate between classical image processing and image deep learning image processing.

      [6]
    2. b)

      How deep learning works for image classification?

      [6]
    3. c)

      This part has equations that don’t show as text. Open the PDF to read it.

      [6]