Exam Cheatsheet

Logical Equivalences:

Logical entailment:

$ \alpha \vDash \beta $ if and only if $(\alpha \lor \lnot \beta)$ is unsatisfiable.

Probability

$$ P(A \mid B) = \frac{P(A \cap B)}{P(B)} $$

$$ P(A \cap B) = P(A \mid B) * P(B) $$

$$ P(A \cup B) = P(A) + P(B) - P(A \cap B) $$

Midterm Review

What is on the assignments is a good representation of the types of exam problems you will see on the exam. These are the kinds of questions that will be asked.

There are also questions covering advantages/disadvantages of certain methods, and when and how they should be used. These questions might be different than what we've seen on the assignments.

Each of the 5-6 topic areas will have 2-3 questions.

Are the following statements tautologies.

Prove something using entailment and resolution.

Translate a sentence into propositional logic

Calculate a probability for a Bayesian Network

Are variables A and B independent.

Compile a joint probability table.

Decision Trees

The syllabus covers the rules of the exam.

The exam will have an exit poll at the exit doors, with two questions:

Everything you have seen so far can be seen, up to "Using Strips"

Preparing for the exam.

The midterm and the final exam, combined, represent 70% of your grade for the course.

A small set of the assignment questions are too difficult, particularly, probability questions in which the format is "Find an example that demonstrates $X$."

Intelligent agents will be covered.

Strips will be covered.


Implication and Equivalence

$x \vDash y \iff (x \implies y) \land (y \implies x)$

$x$ and $y$ are equivalent ($x \equiv y$) if and only if whenever $x$ is true, $y$ is also true, and whenever $x$ is false, $y$ is also false.

The entailment symbol $\vDash$ is not a logical symbol, and is not part of propositional logic.

Bayesian Network

The Markov boundary of a node $A$ in a Bayesian network is the set of nodes composed of $A$'s parents, $A$'s children, and $A$'s children's other parents.

Prove B and C are conditionally independent, given A.

Apply Forward Chaining and Backward Chaining, there

For backward chaining, only one rule applied, such as depth first search, or backtracking, to show the query is entailed.

Forward & Backward Chaining

Forward-chaining algorithm:

Backward-chaining algorithm: *