AIMA Artificial Intelligence a modern approch

AIMA-exercises is an open-source community of students, instructors and developers. Anyone can add an exercise, suggest answers to existing questions, or simply help us improve the platform. We accept contributions on this github repository.

Exercise 14.21

Consider the query ${\textbf{P}}({Rain}{Sprinkler}{true},{WetGrass}{true})$ in Figure rain-clustering-figure(a) (page rain-clustering-figure) and how Gibbs sampling can answer it.

  1. How many states does the Markov chain have?

  2. Calculate the transition matrix ${\textbf{Q}}$ containing $q({\textbf{y}}$ $\rightarrow$ ${\textbf{y}}’)$ for all ${\textbf{y}}$, ${\textbf{y}}’$.

  3. What does ${\textbf{ Q}}^2$, the square of the transition matrix, represent?

  4. What about ${\textbf{Q}}^n$ as $n\to \infty$?

  5. Explain how to do probabilistic inference in Bayesian networks, assuming that ${\textbf{Q}}^n$ is available. Is this a practical way to do inference?

View Answer