CS-601 Machine Learning - May 2022 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 basic design issues and approaches to machine learning?
   How to answer

   • Define machine learning design problem with task, performance, experience
   • List design issues: training data choice, target function representation, function approximation algorithm
   • Describe approaches: supervised, unsupervised, reinforcement
2. 7 Marks
   bDefine statistical theory and how it is performed in machine learning.
   How to answer

   • Define statistical learning theory
   • Explain empirical risk, structural risk, generalization
   • Describe role in model selection and performance in ML
3. 7 Marks
   aDifferentiate between Training data and Testing data.
   How to answer

   • Define training data vs testing data purpose
   • Compare on usage, labeling, role in learning vs evaluation
   • Explain split ratio, overfitting and generalization aspect
4. 7 Marks
   bWhat do you mean by Gradient Descent?
   How to answer

   • Define gradient descent and cost function minimization
   • Explain iterative update rule with learning rate
   • Mention types: batch, stochastic, mini-batch and convergence
5. 7 Marks
   aWrite the algorithm for back propagation.
   How to answer

   • Initialize weights and define feedforward step
   • Compute error and backpropagate delta terms
   • Give weight update rule and stopping criteria
6. 7 Marks
   bExplain in detail principal component analysis for dimension reduction.
   How to answer

   • Define dimensionality reduction need and PCA goal
   • Explain covariance, eigenvectors, principal components steps
   • Mention selection of top-k components and reconstruction
7. 7 Marks
   aWhat is Reinforcement learning? Explain its detailed concepts.
   How to answer

   • Define reinforcement learning and agent-environment loop
   • Explain reward, policy, value function, exploration vs exploitation
   • Give examples and applications
8. 7 Marks
   bExplain Locally weighted linear regression.
   How to answer

   • Define locally weighted regression as non-parametric method
   • Explain weighting kernel and query-point fitting
   • Give cost function and parameter update formula
9. 7 Marks
   aWhat is model selection in Machine Learning?
   How to answer

   • Define model selection problem
   • Explain train-validation-test split and cross-validation
   • Mention bias-variance and selection criteria
10. 7 Marks
    bDescribe the concept of MDP.
    How to answer

    • Define MDP tuple (S,A,P,R)
    • Explain Markov property, transitions and rewards
    • Mention policy and optimal value functions
11. 7 Marks
    aExplain Q learning algorithm assuming deterministic rewards and actions.
    How to answer

    • Define Q-function and deterministic setting
    • Give Q-learning update rule with learning rate
    • Explain algorithm steps and convergence condition
12. 7 Marks
    bExplain the concept of Bayesian theorem with an example.
    How to answer

    • State Bayes theorem formula with prior, likelihood, evidence
    • Explain each term with example calculation
    • Mention use in classification and Bayesian learning
13. 7 Marks
    aWhat is Support Vector Machine (SVM)? Discuss in detail.
    How to answer

    • Define SVM and maximum margin hyperplane
    • Explain support vectors, kernel trick for non-linear data
    • Mention optimization objective and applications
14. 7 Marks
    bDefine Bayesian learning and how it impacts in machine learning?
    How to answer

    • Define Bayesian learning and probabilistic framework
    • Explain prior to posterior updating with data
    • Discuss impact: handling uncertainty, overfitting, prediction
15. 14 Marks
    Explain the following terms with example :

    1. i)Convex optimization in machine learning
    2. ii)Multilayer network
    3. iii)Attention model
    4. iv)Natural Language Processing
    How to answer

    • Define each term: convex optimization, multilayer network, attention, NLP
    • Explain key principle and working with example for each
    • Mention ML applications of each term