Chapter 12. Solving other inference tasks

 

This chapter covers

  • How to query the joint probability of multiple variables
  • How to compute the most likely values of all variables in your model
  • How to compute the probability of your observed evidence

So far, you’ve looked at ways of answering questions such as “What’s the probability that the printer power switch is off, given that it’s not printing correctly?” or “What’s the probability that I’ll win the Oscar, given my skill and the movie’s quality?” All of these questions can be summarized as answering the query of computing the posterior probability distribution over a single variable, given the evidence. A variety of other inference tasks can be performed using probabilistic programming. These include the following:

  • Computing the joint probability distribution over multiple variables
  • Computing the most likely values of the variables, given the evidence
  • Computing the probability of the evidence (the probability that a possible world generated by the model will satisfy the evidence)
  • Monitoring the state of a dynamic system over time
  • Learning the parameters of a model so they can be used for reasoning later

12.1. Computing joint distributions

12.2. Computing the most probable explanation

12.3. Computing the probability of evidence

12.4. Summary

12.5. Exercises