CSC384 Review Checklist
Here is a brief overview of the topics we've covered in 384 to help you
organize your review of material. A couple of topics will not be covered
on the exam as noted. The material you are responsible for includes: relevant
chapters in the text; material from class; online lecture summaries and
notes; any supplemental notes/handouts; and material from assignments.
-
What is AI All About? (Readings: Text Ch.1, online notes)
-
Representation and Reasoning Systems (Readings: Text Ch.2; Ch.3
excluding 3.7; online notes)
-
Representation and Reasoning Systems and Simplifying Assumptions
-
Definite Clause Language: Syntax (facts, rules, KBs)
-
Definite Clause Language: Semantics (interpretations, satisfaction, models,
logical consequences)
-
Queries and Answer Sets
-
Proof Procedures: Soundness, Completeness
-
Bottom Up Proof Procedure
-
Top-Down Proof Procedure (SLD-Resolution)
-
Unifiers
-
Uses of the Definite Clause Language
-
Prolog
-
Building a Knowledge Base
-
Graph Search (Readings: Text Ch.4 excluding "Dynamic Programming"
(which is in 4.6) and excluding 4.7; online notes; please take note that
material on game tree search is not covered in the text, only in the online
notes)
-
Problem Solving as Graph-based Search
-
Formalization of Graph Search
-
Types of Solutions (shortest path, least-cost, satisficing)
-
Generic Search Procedure (frontier, addition to frontier, selection)
-
Search Trees
-
Depth-first Search
-
Breadth-first Search
-
Least-Cost First Search
-
Heuristics
-
Best-first Search
-
A* Search (admissibility, monotone restriction)
-
Iterative Deepening
-
Implicit Search Graphs
-
Island-driven search
-
Game-tree search
-
Planning (Readings: Text Ch.8 excluding "Event Calculus", (which
is in 8.2) and excluding "Partial-order Planning (which is in 8.3); you
should read enough about the "Situation Calculus" (8.2) to be able to follow
the text, but you are not directly responsible for material on the Situation
Calculus; online notes)
-
General Motivation for Planning
-
World Representations: Explicit World Rep'n, Closed World Rep'n, Derived
Relations
-
Action Representations: STRIPS (preconditions, add and delete lists)
-
Planning as Forward Search
-
STRIPS Planning: Basic STRIPS, STRIPS with Goal Protection, Serializability
-
Regression Planning
-
Reasoning Under Uncertainty (Readings: Text Ch.10, excluding "Information
Theory" (which is in 10.2.); online notes). You are not responsible for
material on Dynamic Bayesian Networks from lecture slides.
-
Decision Making: The need to quantify our uncertainty
-
Basics of Probability Theory: Random Variables, Distributions, Measures
(Semantics), Basic Rules
-
Conditional Probability
-
Probabilistic Inference: Conditioning, Computational Bottlenecks
-
Independence and Conditional Independence
-
What Independence Buys Us
-
Bayesian (Belief) Networks: DAGs, CPTs, underlying semantics
-
Constructing a Bayes net
-
D-separation
-
Simple Inference Schemes
-
Variable Elimination
-
Utility Theory: principle of Maximum Expected Utility
-
One-shot versus Sequential Decision Problems: Decision Trees
-
Decision Networks