CSE422: Artificial Intelligence

Taught CSE422: Artificial Intelligence during Summer 2025. This course presents the foundational mathematical principles, search algorithms, knowledge representations, and machine learning strategies that enable computational systems to exhibit intelligent behavior.

Pre-requisite: CSE221
Semester Taught: Summer 2025
Institution: BRAC University, Department of Computer Science and Engineering

Course Description

This course provides a comprehensive introduction to Artificial Intelligence. Students explore intelligent agent paradigms, state-space formulations, problem-solving through uninformed and heuristic search algorithms (including A* and local search), adversarial game playing (Minimax with Alpha-Beta pruning), constraint satisfaction problems (CSPs), knowledge representation and reasoning, probabilistic models under uncertainty (Bayesian networks), and core principles of machine learning. The curriculum emphasizes rigorous algorithmic formulation, complexity analysis, and practical implementation.

Course Objectives

• Introduce definitions of artificial intelligence, rationality, task environments (PEAS), and agent architectures.
• Analyze state-space representations and search strategies for informed and uninformed problem solving.
• Master adversarial search algorithms and heuristic evaluation functions for competitive multi-agent game playing.
• Formulate, analyze, and solve Constraint Satisfaction Problems (CSPs).
• Study probabilistic reasoning under uncertainty and fundamental machine learning paradigms.

Course Topics & Outline

• Intelligent Agents: Environments, agent types (simple reflex, goal-based, utility-based), and state-space problem formulation.
• Uninformed Search Algorithms: Breadth-First Search (BFS), Depth-First Search (DFS), Depth-Limited Search, Iterative Deepening, and Uniform Cost Search (UCS).
• Informed (Heuristic) Search: Greedy Best-First Search, A* Search, admissibility and consistency of heuristics, memory-bounded search (IDA*).
• Local Search & Optimization: Hill Climbing, Simulated Annealing, Local Beam Search, and Genetic Algorithms.
• Adversarial Search & Game Playing: Two-player zero-sum games, Minimax decision algorithm, Alpha-Beta pruning, and evaluation function design.
• Constraint Satisfaction Problems (CSPs): Constraint propagation, Backtracking search, Forward Checking, Arc Consistency (AC-3), and MRV/degree heuristics.
• Knowledge & Logic: Propositional logic, inference rules, First-Order Logic (FOL), and knowledge-based agents.
• Uncertainty & Probabilistic Reasoning: Probability theory, conditional independence, Bayes’ rule, and Bayesian Networks.
• Machine Learning Foundations: Supervised vs. unsupervised learning, classification, linear/logistic regression, and decision trees.

Textbooks & Reference Materials

1. Artificial Intelligence: A Modern Approach, Stuart Russell and Peter Norvig, 3rd/4th Edition, Pearson Education, ISBN: 9780136042594.
2. Departmental lecture slides, problem sets, programming assignments, and code materials provided via buX.