Quiz dossier / Deep Learning Book

19 - Approximate Inference

Advanced mastery of Chapter 19, Approximate Inference, covering every named section across book pages 631-653.

11 questionsEstimated 90 min0 flagged

Session status

Draft

Question 1

3 point(s)

Inference as optimization

How can an intractable posterior $p(h|v)$ be approximated?

Choose a tractable

q(h)

and optimize a divergence or bound measuring its mismatch to the posterior

Assume every latent state is equally likely without checking

Delete the observed variables

Replace probability with an arbitrary unnormalized constant