One quiz will be on Wednesday, January 24th for the undergraduate class, and the other on Thursday the 25th for the graduate class. The quizzes are closed-book, closed notes. Below is a list of questions, one for each chapter. This set of questions can also serve to give you some guidance as to the way you'll need to study the Overview assignment and to what depth. The Overview lecture notes themselves are here.
For this quiz and for other tests in this and your other courses, you should practice writing out short answers to the questions. Note that simply reading the questions and the book is not the best preparation for the quiz. As I've told students many times, on a test, I don't grade your reading, I grade your writing!
Grad students: The questions below are based on the sections of the chapters assigned for the undergraduate course. You may be asked a few questions that are based on the additional chapter sections that you have been assigned for the Overview.
Chapter 1: What are some problems with the "laws of thought" approach (thinking rationally) as compared to the rational agent approach (acting rationally)?
Chapter 2: Give a definition of a rational agent with an additional sentence interpreting the definition.
Chapter 3: List and briefly describe the four components of a problem as specified in constructing a search for solutions by an agent.
Chapter 4: Describe one or two of the common heuristics used to reduce search costs.
Chapter 5: Using the example of map coloring from chapter 5, describe the concept of constraint satisfaction.
Chapter 6: Describe either pruning or evaluation functions as used in designing systems for games such as chess.
Chapter 7: Describe how the following task environment components are realized in formulating the Wumpus World problem: Performance measure, environment, actuators, and sensors.
Chapter 8: Describe the notions of objects, relations, and functions that allow first-order logic to go beyond propositional logic.
Chapter 9:
The rule of Universal Instantiation can be applied to axioms such
as
∀x King(x) ∧ Greedy(x) ⇒ Evil(x)
Give an example of the type of result this could produce.
Chapter 10: Give a few examples of AbstractObjects and GeneralizedEvents from Fig. 10.1. By examples, I mean items further down the tree in the figure.
Chapter 11: Describe the notions of preconditions, effects in planning. Effects are further divided into an add and a delete list - what are they?
Chapter 12: Discuss the notion that in some plans, speeding up a particular action may not speed up the total process. Give a small example, such as getting a bus station an hour early may not get you to your destination any earlier.
Chapter 13: In connection with uncertainty, discuss one or two of the reasons that first-order logic can fail to cope with some domains (pgs. 463-464).
Chapter 14: If I give you the diagram of Fig. 14.2 without the CPTs, explain why the CPT for Alarm has the entries it does.
Chapter 15: A first-order Markov process has a simple relation between the state at a given time, Xt, and the states at all previous times. Describe this simple relationship, using the standard notation if you can.
Chapter 16: Why is utility added to decisions, beyond the computation of the probabilities of outcomes?
Chapter 17: Given the left-hand portion of Fig. 17.2, minus the arrows, fill in the arrows (the policy) and comment on what they mean, especially the choice of arrows in the vicinity of the negative reward (-1) state.
Chapter 18: Explain the notion of induction from examples in learning, e.g., learning from a table such as Fig. 18.3.
Chapter 19: Name and describe any one of the techniques of knowledge-based learning that are briefly described in Sec. 19.6.
Chapter 20: No questions will be asked about this chapter.
Chapter 21: In a sentence or two, describe reinforcement learning.
Chapter 22: The syntactic rule, S → NP VP contains non-terminals. In contrast to these, give a couple of examples of terminals.
Chapter 23: Discuss why the generation of a word sequence by a trigram model begins to look like legitimate bit of natural language, as compared to using a unigram model.
Chapter 24: Compare feature-based and model-based approaches to perception (primarily, visual perception).
Chapter 25: What, roughly, is meant by "degrees of freedom" in robot effectors?
Chapter 26: What is the difference between the questions, "Can machines act intelligently?" and "Can machines really think?".
Chapter 27: What does the book have to say about the question: "What if AI does succeed?"
Go to CSU520 home page.,
to CSG120 home page.,
RPF's Teaching Gateway, or
homepage