Text: Artificial Intelligence – A Modern Approach 2nd Edition, Russell and Norvig (RN) Available from the campus bookshop.
We begin this course by discussing the notion of an agent intelligently perceiving, reasoning, learning, planning and interacting with its environment. The course then discuss a variety of techniques from areas such as logic, probability and computational learning theory that can be used to implement an intelligent agent. The agent typically in A.I. has been a robot and the environment the physical world. However, new application areas such as software agents (softbots) navigating through the WWW.
Introduction to Artificial Intelligence and the Agent Approach to A.I.
Lectures :3
Chapters 1,2 (RN)
A brief history of A.I., the structure and parts of an agent, the notion of rational behavior, different types of environments require different techniques.
Problem Solving
Lectures :6
Chapters 3,4,5,6 (RN)
Problem solving as search: breadth, depth-limited, bi-directional search, heuristic search techniques: hill climbing, constraint satisfication problems: backtracking search. Decision strategies in game theory.
PAC Learning, Reinforcement Learning and Neural Networks
Lectures :6
Chapters 18,20, 21 (RN)
Decision trees, neural networks, reinforcement learning, computational learning theory
Knowledge Representation and Reasoning
Lectures :6
Chapters 7,8,9,14 (RN)
Propositional logic, propositional inference, first-order logic, first order inference, unification, lifting, chaining, resolution, theorem provers, Bayesian belief graphs.
Planning and Natural Language Processing
Lectures :3
Chapters 11, 22 (RN)
GraphPlan algorithm, formal grammars, semantic interpretation, (dis)ambiguity, grammar induction
Class Participation: 10%
Assignment #1: 20%
Assignment #2: 20%
Final Exam: 50%