Skip to content
ml-recall

#lower-bound

1 question from Autoencoders & VAEs

Timed test

1 question · 1 min · score and explanations at the end

Take a test
Q 1 / 1Score 0 / 0

Question 1: Single choice

hardAI-checkedAutoencoders & VAEs

A distribution p(x)=p~(x)/Zp(x) = \tilde{p}(x) / Z has an intractable normaliser ZZ. Variational inference picks a tractable qq and minimises J(q)=∑xq(x)log⁡q(x)p~(x).J(q) = \sum_x q(x) \log \frac{q(x)}{\tilde{p}(x)}. What does J(q)J(q) tell you about log⁡Z\log Z?

Choose one option