Artificial Intelligence and Machine Learning

EC25401:

Artificial Intelligence and Machine Learning

Course objective:

To provide foundational AI/ML knowledge covering core techniques and enable real-world application using modern tools and frameworks.

Introduction To AI

Artificial Intelligence: Introduction, Applications, Problem types. Problem formulation, Intelligent Agents: Types, Architectures, and Environments, PEAS framework. Reasoning and Logic: Propositional Logic, First-order Logic, Inference techniques including Forward and Backward Chaining

Search Strategies

Solving problems by searching, Design issues in search algorithms. Uninformed Search: BFS, DFS. Heuristic Search: Generate-and-Test, Hill Climbing, Best-First Search, A*, Alpha-Beta pruning. Advanced methods: Problem Reduction, AO* Algorithm, Constraint Satisfaction, Means-Ends Analysis.

Artificial Neural Networks

Introduction to Neural Networks, Activation Functions, Optimization Algorithms: Gradient Descent. Architectures: Perceptrons, Adaline, Multilayer Perceptrons. Training: Backpropagation, Procedures, Network Tuning.

Machine Learning Basics

Machine Learning: Basics, Applications. Comparison with Data Mining and Big Data Analytics. Supervised Learning: Naive Bayes, KNN, Decision Trees. Unsupervised Learning: k-means clustering. Introduction to Reinforcement Learning.

Forecasting, Advanced Learning & Ensemble Methods

Forecasting: Non-linear and Logistic Regression, Random Forests, Bayesian Networks, Bias-Variance Tradeoff, Tuning and Model Selection. Clustering: EM Algorithm, Hierarchical Clustering, Post-Clustering Supervised Learning. Kernel Methods: SVM, Soft Margins, Kernel Trick. Ensemble Methods: Bagging, Boosting, Stacking, AdaBoost, Cross-Validation. Dimensionality Reduction: PCA, LDA, MDS, Feature Selection Techniques.

Suggested Experiments:

1. Implementation of a simple intelligent agent simulation

2. Solving AI problems using BFS and DFS algorithms

3. Implementing heuristic search strategies (A*, Hill Climbing)

4. Design and train a perceptron for basic classification

5. Building a multi-layer neural network using backpropagation

6. Data preprocessing and feature selection for ML tasks

7. Apply supervised learning algorithms (Naive Bayes, Decision Tree, KNN)

8. Unsupervised learning using k-means and Apriori algorithms

9. Regression techniques and Random Forest model implementation

10. Dimensionality reduction using PCA and LDA with visualization

Weightage: Continuous Assessment: 50%, End Semester Examinations: 50%

Assessment Methodology: Quiz (5%), Project (15%), Assignment Programs (25%),

Practical (25%), Internal Examinations (30%)

References:

1. Stuart Russell and Peter Norvig, Artificial Intelligence: A Modern Approach, 4th Edition, Pearson, 2021.

2. Tom M. Mitchell, Machine Learning, McGraw Hill Education, 2017.

3. Ian Goodfellow, YoshuaBengio, and Aaron Courville, Deep Learning, MIT Press, 2016.

4. Pattern Recognition and Machine Learning. By Christopher M. Bishop, Springer, 2006.

5. Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow. By

AurélienGéron, O'Reilly Media, 2022