Chapter 7. Approximate Inference Methods

In the previous chapter, we have learned that as the tree width of a graphical model increases, the exact inference becomes infeasible. The motivation to pursue approximate inference comes from real-world networks where the exact inference is intractable.

In this chapter, we will learn about the methods to calculate approximate inference. We will revisit message passing algorithms mentioned in the previous chapter and learn how they can also be used to calculate approximate inference when the network is not tree structured. We shall explore inference from the perspective of optimization and learn about sampling methods. We will also look at some code samples to understand how algorithms implement the approximate ...

Get Building Probabilistic Graphical Models with Python now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.