SUK · Third Year B.Tech. Computer Science and Engineering (Artificial Intelligence and Data Science) · NEP 2020
Data Science Optimization Techniques
AI&DS3205 · Elective III · Semester VI · 3 credits · ISE 10 + mid-semester 30 + ESE 60
One of the Elective III options.
Unit 1: Introduction to Optimisation
6 hoursWhat optimisation is and its role in data science and AI; types of problems; objective functions, constraints and convexity.
Examined in the 60-mark end-semester paper, and in the mid-semester exam if it is taught by then
- Formulating optimisation problems
- Classification of optimisation algorithms
- Convex and quadratic programming
Unit 2: Classical Optimisation Techniques
6 hoursLinear programming: the graphical and simplex methods, duality and sensitivity analysis; integer programming; non-linear programming basics.
Examined in the 60-mark end-semester paper, and in the mid-semester exam if it is taught by then
- Linear programming and duality
- Simplex and revised simplex methods
- Integer linear programming
Unit 3: Gradient-Based Optimisation Methods
6 hoursUnconstrained optimisation; gradient, stochastic and mini-batch gradient descent; Newton and quasi-Newton methods; convergence analysis.
Examined in the 60-mark end-semester paper, and in the mid-semester exam if it is taught by then
- Gradient descent variants
- Newton's and secant methods
- Projected gradient methods
Unit 4: Heuristic and Metaheuristic Optimisation
6 hoursGenetic algorithms, particle swarm optimisation, simulated annealing and ant colony optimisation compared.
Examined in the 60-mark end-semester paper, and in the mid-semester exam if it is taught by then
- Genetic algorithms
- Particle swarm optimisation
- Simulated annealing
- Ant colony optimisation
Unit 5: Optimisation in Machine Learning
6 hoursLoss functions and their optimisation; regularisation; optimisation in neural networks and back-propagation; hyperparameter tuning.
Examined in the 60-mark end-semester paper, and in the mid-semester exam if it is taught by then
- Loss functions in machine learning
- Ridge and lasso regularisation
- Optimisers: AdaGrad, RMSProp and Adam
- Hyperparameter tuning with grid search
Unit 6: Advanced Optimisation Techniques
6 hoursEvolutionary strategies; Bayesian optimisation; multi-objective optimisation; distributed and parallel optimisation; recent research trends.
Examined in the 60-mark end-semester paper, and in the mid-semester exam if it is taught by then
- Differential evolution
- Bayesian optimisation
- Multi-objective optimisation and Pareto fronts
- Distributed data parallel training
Unit contents are summarised from the official SUK syllabus . AiCanCode.org is not affiliated with Shivaji University, Kolhapur.