Final Exam Guide
Thursday, November 17 - Friday, November 18
Exam Information and Format
- The final exam will be conducted online using the D2L Quiz mechanism. The format and the types of questions on the exam will be similar to the questions on the quizzes (including multiple choice, multi-select, short answer, etc.).
- The exam will be available starting on Thursday, November 17, 12:01 AM and will close on Friday, November 18, 11:59 PM. You will have 3 hours to complete the exam from the time you start the exam within the allowable period.
- You may attempt the exam only once.
- The exam will cover all material presented on the Schedule Page of the Class Web site through November 2. See the detailed guide below.
- The exam questions will be very similar to the questions on the quizzes and the examples given in class presentations. Going over these questions and examples would be a good way to prepare for the final exam.
The following material will be covered in the exam
-
AI &
Intelligent Agents
- Definitions and applications of AI
- PEAS description for agents
- Architectural patterns for agent design (reflex agents, goal-based and utility agents, etc.)
- Environments and their properties
-
Problem Solving
and Search
- State-space representation and state-space graphs
- Role of operators and goals (e.g., impact of operators on branching factor in the search space, etc.)
- Problem types and their characteristics (single-state, multiple-state, contingency, etc.)
- Search strategies (depth-first, breadth-first, iterative deepening, uniform cost)
- Analysis of completeness, optimality, and space/time complexity for various search strategies
-
Informed Search
- Admissible heuristics and comparison of various heuristic functions
- Hill-climbing search strategies
- Greedy Best-first search
- A* search
-
Adverserial
Search and Games
- Heuristics in games
- Minimax procedure
- Alpha-Beta pruning
- Expectiminimax algorihtm
- Constraint Satisfaction Problems
- Basics and examples of CSPs
- Constraint Graphs
- Backtracking search for CSPs
- Forward checking and constraint propagation
- Arc consistency
- Ordering strategies such as Min. Remaining Values and Least Constraining Value heuristics
- Local search for CSPs
- Tree structured CSPs
-
Logical
Reasoning
- Logical entailment and logical equivalence in Propositional Logic
- Models and interpretations in Propositional Logic
- Representation of problems in First-Order Predicate Calculus (e.g., translation from English to FOPC)
- Models and interpretations in FOPC, and the semantics of FOPC statements
- Using the rules of inference in proofs (derivations) in Propositional Logic and FOPC
- Conversion of First-Order logic formulas into Conjunctive Normal From (CNF) or clausal form (including "Skolemization")
- Resolution rule of inference and refutation procedure
- Forward and backward chaining and AND-OR proof trees
- Basic Prolog programming
- Reasoning with Uncertainty
- Axioms of probability theory
- Joint probability distributions and making basic inferences
- Bayes’ Rule and its use in making probabilistic inferences
- Combining evidence (Bayesian updating)
- Notion of conditional independence
- Bayesian networks, including making inferences given a BN
- Machine Learning
- Basic concepts in classification and prediction, including evaluation metrics
- Naïve Bayesian classification
- Decision tree classification (including the ID3 decision tree induction algorithm)
- Distance and similarity measures (such as Euclidean distance, Cosine similarity, and Pearson correlation)
- K-nearest-neighbor (KNN) strategy for classification or prediction
- Clustering (including the K-Means algorithm)
Textbook Material Related to the Exam
- Chapter 1 (all)
- Chapter 2 (all)
- Chapter 3 (all)
- Chapter 4 (Section 4.1)
- Chapter 5 (Sections 5.1-5.5)
- Chapter 6 (all)
- Chapter 7 (Sections 7.1-7.5)
- Chapter 8 (Sections 8.1-8.3)
- Chapter 9 (all - through "Example Proofs" subsection of 9.5)
- Chapter 12 (all)
- Chapter 13 (Sections 13.1-13.3)
- Chapter 19 (Sections 19.1-19.4)