Artificial Intelligence: final assessment

This covers the whole course. Take it after the last module.

Final quiz

1. In PEAS, what does the S stand for?
2. Which data structure makes a search breadth-first?
3. What does local search keep in memory?
4. What is a solution to a constraint satisfaction problem?
5. How many rows does a truth table with 3 symbols have?
6. What does A* return when its heuristic is admissible?
7. Why does forward checking help backtracking search?
8. Which method is goal-driven?
9. What does normalizing a set of weights do?
10. What is the probability of a complete assignment in a Bayesian network?
11. In value iteration, what is the best action in a state?
12. Why does arc consistency not always solve a CSP?
13. Why does alpha-beta search deeper with good move ordering?
14. A DPLL solver sets a variable without branching. Why?
15. Seeing wet grass and then learning it rained makes the sprinkler less likely. What is this called?
16. Why is Q-learning called off-policy?

Practical assessment: Build an agent for a grid world

Without using hints, write a program that reads a grid map and solves it three ways: the shortest path with A*, the optimal policy when moves slip with value iteration, and a learned policy with Q-learning.

Final project

A robot that plans, tracks and learns: Build a grid-world robot that finds paths with A*, keeps track of its position from noisy sensor readings, plans under slippery moves with value iteration, and learns the same task with Q-learning.

How it is evaluated

To complete this course