B.TechSemester 72023-24Optimization In Machine LearningKAI071
View PDFDownload

Optimization In Machine Learning (KAI071) - AKTU Question Paper 2023-24

B.Tech · Semester 7 · Free PDF Download

This is the official AKTU Optimization In Machine Learning Previous Year Question Paper for B.Tech Semester 7, academic session 2023-24. Published by Dr. A.P.J. Abdul Kalam Technical University (AKTU/UPTU), Lucknow. Free PDF download — no login required.

Course:B.Tech
Semester:Semester 7
Session:2023-24
University:AKTU / UPTU

Rate this paper

Questions Asked in 2023-24

Optimization In Machine Learning (KAI071) — extracted question text · 100 marks · 3 Hours

Section AAttempt all q u e s t i o n s i n b r i e f . 2 x 10 = 20
  • a
    How does convexity play a cruci al role in optimization problems? 2
  • b
    Illustrate the application of c onvex optimization in real-world scenarios. 2
  • c
    Explain Nesterov's applicati on in convex optimization. 2
  • d
    Investigate Moreau–Yos ida regularization. 2
  • e
    Explain the regularization process. 2
  • f
    Explain the concept of dual d ecomposition in optimization 2
  • g
    How Douglas–Rachford splitti ng addresses challenges? 2
  • h
    How do optimization algorithms strat egically navigate saddle points? 2 i. What are the implications for c onvergence and optimization efficiency? 2 j. How does the structure of the optimization landscape impact th e choice of optimization algorithms?
Section BAttempt any three o f t h e f o l l o w i n g : 10 x 3 = 30
  • a
    Compare and contrast linear p rogramming, second-order cone p rogramming, and semidefinite programming. Provide real-world examples where each type of convex program is applicable
  • b
    Explain the concept of duality in convex optimization. How d oes duality provide insights into the primal and dua l aspects of an optimization pr oblem? Discuss the relationship between the primal and dual solutions
  • c
    Compare and contrast mirror de scent and traditional gradient descent methods. Highlight the advantages and disadvantages of mirror descent, a nd provide a real-world example where mirror descent might outperform other optimization algorithms
  • d
    Compare the convergence proper ties of Augmented Lagrangian m ethods and ADMM. Discuss scenarios where one method might be preferred over the other based on the nature of the optimization problem
  • e
    Compare and contrast Polyak–Juditsky averaging with other me thods used in stochastic gradient descent. H ow does it contribute to the conv ergence and stability of optimization algorithms, especially in the context of deep learning?
Section CAttempt any one p a r t o f t h e f o l l o w i n g : 10 x 1 = 10
  • a
    Describe the Karush-Kuhn-Tucker (KKT) conditions in the cont ext of convex optimization. How do these condi tions characterize optimality i n convex programs? Provide examples to illustrate the application of KKT conditions
  • b
    Provide detailed explanations and examples of different conv ex programs. Highlight the specific mathemat ical formulations and problem st ructures for each type of convex program
  • a
    Elaborate on the Frank–Wolfe method and its applications in constrained optimization. Discuss how the method addresses challenges posed by large- scale optimization problems and p rovide a scenario where it is particularly effective
  • b
    Explore the concept of Ordinary Differential Equations inter pretations in the context of optimization algorithms
  • a
    Discuss the role of dual methods in optimization and their a pplications in solving primal-dual problems. Pro vide an example illustrating t he use of dual methods and explain their advanta ges over primal methods in certain scenarios
  • b
    Provide an in-depth overview of proximal gradient methods an d their role in handling non-smooth and non-convex optimization problems
  • a
    Compare and contrast the Alternating Direction Method of Mul tipliers with other optimization algorithms, highlighting its strengths and weaknesses
  • b
    Describe the Douglas–Rachford splitting algorithm and its ap plication in solving convex optimization problems. Discuss situations where this algorithm is particularly effective and the conditions under which it converges
  • a
    Elaborate on the application of Langevin dynamics in Bayesia n inference. How does it relate to escaping saddle points and contribute to efficient sampling in high-dimensional spaces, particularly in the context of deep learning models?
  • b
    Examine the Bayesian interpr etation of Langevin dynamics. Ho w does Langevin dynamics connect to the posterior distribution in Baye sian models, and what are the implications f or exploring the posterior space in the context of probabilistic modeling?

Question text is extracted from the official AKTU question paper PDF above. Check the original PDF for equations, diagrams and complete wording. Extracted: 2026-08-23.

More B.Tech Semester 7 (2023-24) Papers

Other subjects from same semester and session

Distributed Computing System2023-24Machine Learning2023-24Power Plant Engineering2023-24Soil Water Conservation Engineering2023-24Operations Research2023-24Text Analytics And Natural Langugae Processing2023-24Vision For Humane Society2023-24Renewable Energy Resources2023-24
View all B.Tech Semester 7 2023-24 papers →

Syllabus & More PYQs

Paper solve karne se pehle unit-wise syllabus dekh lo

AKTU 4th Year PYQs — All Branches