CS 2400, Hasker's Section: Final Exam Review Notes
- Explain how a neural network illustrates the PEAS model of agents as
described in note 1.
- Explain how mini-max search illustrates the PEAS model of agents as
described in note 1.
- Outside of problems considered in this course, give an application of
A* search in automobile repair, retail, or restaurants.
- Outside of problems considered in this course, give an application of
neural networks in automobile repair, retail, or restaurants.
- Characterize when A* is a better solution than neural networks.
- Characterize when neural networks are better at solving problems than A*.
- What is the problem of "noisy" domains? How would a noisy domain
impact A*? How would it impact a neural network?
- Mini-max search was generally discussed in the context of game
playing. Give a different scenario for applying mini-max search.
- Which of the AI algorithms discussed this term would be the most
helpful for writing music? Defend your answer.
- How do we solve problems using predicate logic?
- It is likely that all of the AI techniques we have discussed this
term would be used when driving a car. Explain how A*, logic,
neural networks, and genetic algorithms might play a
role. Do not consider route planning in your answer - assume the car is
given directions by an existing navigation system.
- Which of the AI methods discussed this term would you use in a
robotic mouse used to entertain a cat? Its behavior should be reasonably
close to a real mouse with the difference that it would take more risks.
- How would you apply Q-learning to train a heuristic for use in A*?
Assume the domain is to construct washing machines efficiently.
- Specific problems were used to illustrate each of the following
problems. Explain why that problem was either a good fit for the
algorithm or not the recommended solution for real projects:
- A*: path finding
- Neural networks: tic-tac-toe
- Q-Learning: finding a path through a house
- Genetic algorithms: traveling salesman problem