CY-701 Machine Learning - Nov 2023 Question Paper

Under each question, How to answer lists the points an answer should cover, outlined by CampusPrep from the syllabus. Check the details against your textbook.

1. 7 Marks
   aWhat are the major limitations of machine learning?
   How to answer

   • Define scope of ML limitations
   • Data dependency, quality and bias issues
   • Generalization, overfitting and interpretability limits
   • Computational cost and deployment challenges
2. 7 Marks
   bDefine machine learning. Discuss regression analysis in detail.
   How to answer

   • Define machine learning with types
   • Define regression analysis
   • Explain linear / logistic regression steps
   • Applications and assumptions
3. 7 Marks
   aWhat is the gradient descent algorithm in ML? Explain.
   How to answer

   • Define gradient descent and cost function
   • Explain iterative update rule with learning rate
   • Describe batch, stochastic, mini-batch variants
   • Convergence and challenges
4. 7 Marks
   bWhat is the purpose of sigmoid activation function in Machine Learning?
   How to answer

   • Define sigmoid function and formula
   • Explain squashing to (0,1) for probability
   • Role in logistic regression and binary classification
   • Limitation like vanishing gradient
5. 10 Marks
   aDiscuss the implementation of Convolutional Neural Network in Tensor Flow.
   How to answer

   • Outline CNN layers: convolution, pooling, fully connected
   • TensorFlow workflow: data pipeline, model building
   • Training with compile, fit, loss and optimizer
   • Evaluation and example code structure
6. 4 Marks
   bWrite a short note on the concept of padding.
   How to answer

   • Define padding in CNN
   • Explain same vs valid padding
   • Purpose to preserve spatial dimensions
   • Effect on border information
7. 8 Marks
   aBriefly explain the concept of Recurrent Neural Network (RNN).
   How to answer

   • Define RNN and recurrent connections
   • Explain unfolding through time and memory
   • Training with BPTT
   • Applications and vanishing gradient issue
8. 6 Marks
   bGive an overview of Q-Learning.
   How to answer

   • Define Q-learning as model-free RL
   • Explain Q-table, states, actions, rewards
   • Q-value update formula
   • Exploration vs exploitation
9. 8 Marks
   aWhat is Support Vector Machine? Discuss how Linear Support Vector machine works?
   How to answer

   • Define SVM and hyperplane, support vectors
   • Explain margin maximization
   • Linear SVM formulation and working
   • Hinge loss and classification
10. 6 Marks
    bList and explain application of machine learning in computer vision.
    How to answer

    • List CV tasks: classification, detection, segmentation
    • Explain role of ML/CNN features
    • Examples: face recognition, medical imaging
    • Pipeline and challenges
11. 8 Marks
    aExplain the difference between linearity vs. non-linearity model.
    How to answer

    • Define linear vs non-linear models
    • Decision boundary and mathematical form
    • Comparison table of assumptions and examples
    • When to use each
12. 6 Marks
    bExplain the importance of Data Preprocessing in Machine Learning.
    How to answer

    • Define data preprocessing
    • Explain cleaning, transformation, reduction
    • Impact on accuracy and training
    • Examples: handling missing values, scaling
13. 7 Marks
    aDiscuss the concept of Dimensionality Reduction.
    How to answer

    • Define dimensionality reduction
    • Need: curse of dimensionality
    • Techniques: PCA, LDA, t-SNE
    • Benefits and information loss trade-off
14. 7 Marks
    bWrite the difference between Q-Learning algorithm and SARSA algorithm.
    How to answer

    • Define Q-learning off-policy vs SARSA on-policy
    • Update rules comparison
    • Exploration behavior difference
    • Tabular comparison with examples
15. 14 Marks
    Explain any two of the following :

    1. i)Steps in Natural Language Processing
    2. ii)Batch Normalization
    3. iii)Convex Optimization
    4. iv)Flattening in CNN
    How to answer

    • Select any two topics from i-iv
    • Define each with key steps/concepts
    • Explain working with diagram/flow
    • Applications and significance