SIVI models can be trained with a reverse-KL path gradient whose intractable score is estimated by importance sampling with a conditional normalizing flow proposal learned via forward KL.
Stabilizing training of affine coupling layers for high-dimensional variational inference
1 Pith paper cite this work, alongside 1 external citations. Polarity classification is still indexing.
1
Pith paper citing it
1
external citations · OpenAlex
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Revisiting Unbiased Implicit Variational Inference
SIVI models can be trained with a reverse-KL path gradient whose intractable score is estimated by importance sampling with a conditional normalizing flow proposal learned via forward KL.