CS 430: Survey of AI - Quantifying Uncertainty

These flashcards cover key concepts discussed in the lecture on quantifying uncertainty in AI, including decision theory, Bayes' rule, independence, and Naïve Bayes models.

What is an example of uncertainty in the real world as discussed in CS 430?

If I leave for SFO 60 minutes before my flight, will I be there in time?

What does probabilistic assertions summarize?

Effects of ignorance and laziness.

What does the decision theory combine?

Probability theory and utility theory.

What does the Product Rule express?

P(a | b) = P(a, b) / P(b) for conditional distributions.

In the Chain Rule, how is a joint distribution expressed?

As a product of conditional distributions: P(x1, x2, x3) = P(x3 | x1, x2) P(x1, x2).

What is the aim of Inference by Enumeration?

To compute probabilities such as P(W) or P(W | winter).

What does Bayes' Rule allow us to do?

Build one conditional from its reverse; it describes an update from prior to posterior probability.

What does it mean for two variables X and Y to be independent?

P(x, y) = P(x) P(y) for all x, y.

What is conditional independence?

X is conditionally independent of Y given Z if P(x | y, z) = P(x | z).

What does the Ghostbusters model illustrate?

Using sensor readings to infer the location of a ghost in a grid.

What is a Naïve Bayes model?

A model where one discrete query variable is conditionally independent of all other variables given the query variable.

How is Naïve Bayes applied in text classification?

By determining the category of a text based on the occurrence of key words and their probabilities.

What do the probabilities P(Category = negative) and P(Category = positive) represent?

The prior probabilities for the negative and positive categories respectively.

How do you classify new text items using Naïve Bayes?

By checking the occurrence of words and using them as evidence to classify.

What probability indicates that a sentence is likely negative in Naïve Bayes classification?

The higher probability for the negative category computed from the occurrence of words.

What is a practical example of using probabilistic assertions?

Determining the credibility of a witness in a trial based on their previous testimony and reliability.

How does utility theory relate to decision making?

Utility theory helps assess the satisfaction or value derived from different choices, influencing the decision-making process.

State the formula for conditional independence.

X is conditionally independent of Y given Z if P(x | y, z) = P(x | z).

Fill in the blank: In decision theory, choices are made based on maximizing _.

expected utility.

What are the limitations of the Naïve Bayes classifier?

Assumes feature independence and may not perform well when features are correlated.

What does it mean when a variable is marginally independent?

Marginal independence indicates that the probability of the variable is unaffected by the presence of another variable.

How can Bayes' Rule be applied in spam filtering?

By calculating the probability of an email being spam based on the presence of certain words.