CIS 412: Artificial Intelligence
Midterm Exam Outline
Exam Date: We 03/06/2024
Time: 3:00PM - 4:15PM (75 minutes)
Location: Textile Building (TEX) 102
Only a pen or pencil is allowed. NO tech of any kind is permitted. This includes (but is not
limited to) laptops, tablets, cell phones, smart watches, and calculators. NO notes, textbook,
cheat sheets, or any other reference materials are permitted. The exam is individual.
You will NOT be asked for pseudocode. However, you should understand the agent types and
algorithms enough that you can explain or walk through them like we do in-class. For concepts,
you will NOT be asked to recite definitions of terms, but the questions will require you to know
the definitions of terms in order to answer them correctly. The exam will cover weeks 1-6.
A midterm review session will be held in-class the Monday before the exam (March 4th).
Below is a complete list of topics that may appear on the exam.
Week 1 - Introduction to AI
- What is AI? (from the Acting Rationally perspective)
- Perception-Action Problem
Week 2 - Intelligent Agents
- Perception-Action Problem (continued from Week 1)
- Rational Agents
- Describing Task Environments: PEAS
- Categorizing Task Environments (using the 6 dimensions)
- Types of Agents:
- Simple Reflex
- Model-Based Reflex
- Goal-Based
- Utility-Based
Week 3 - Search (Uninformed)
- Defining a Problem
- Searching for Solutions
- Tree vs. Graph Search
- Uninformed Search Strategies:
- Breadth-First Search
- Depth-First Search
- Uniform-Cost Search
- Evaluate Search Algorithms✝
Week 4 - Search (Informed)
- Informed Search Strategies:
- Best First Search
- A* Search
- Heuristics (admissible, dominant, and consistent)
- Evaluate Search Algorithms✝
- Local Search Strategies (just the general idea of these):
- Hill Climbing Search
Week 5 - Adversarial Search
- Defining a Game as a Search Problem
- Minimax
- α-β Pruning
Week 6 - Logical Agents
- Knowledge-Based Agents (TELL / ASK)
- The Wumpus World
- Models & Entailment
- Logical Inference Algorithms:
- Forward Chaining
- Backwards Chaining
- Both require a good understanding of Horn clauses
✝ You don’t need to memorize complexities, but you should have a general understanding of the
completeness, optimality, time and space complexities of each search algorithm. For example,
you should know which algorithms are fastest, most efficient, etc.
(symbol = dagger)