MIStatlE
中文

← Knowledge mapψ Learning Systems

Generative Models

5 concepts, 0 written; shares concepts or relations with 5 other subjects. Earlier ideas sit to the left, a region holds one chapter, and relations that leave the subject hang on the cards as tags.

12345 earlierlater Score-based generationψGenerative Models applied in extends to extends to extends to KL Divergence &Cross Entropy VAE & ELBO Score Matching ConditionalTechnique ProbabilityFlow ODE Shannon Entropy → KL Divergence & Cross Entropy · extends to — KL divergence is relative entropy: entropy extended from one distribution to a pair.Information Theory · Shannon E… → VAE & ELBO ↔ LeWorldModel / SIGReg · contrast — Both learn latent representations: the VAE reconstructs observations, LeWorldModel predicts only in latent space and relies on SIGReg against collapse.→ Deep Learning · LeWorldModel…
Relationsposition is order: read along the rail, left to righta region holds one chaptera tag on a card leads to another subject
NodesConceptMethodModelwrittenplanned0 / 5 written

Select a concept to see its summary and relations0 / 5 written

Score-based generation0/3

Other0/2

Select a node to see its summary and how it relates to the others.

Relations in this subject, in words6

  • Shannon Entropyextends toKL Divergence & Cross Entropy

    KL divergence is relative entropy: entropy extended from one distribution to a pair.

  • KL Divergence & Cross Entropyapplied inVAE & ELBO

    The regularizer in the ELBO is a KL divergence.

  • VAE & ELBOextends toScore Matching

    From a likelihood bound for latent-variable models to objectives that learn the score directly.

  • Score Matchingextends toConditional Technique

    Conditioning makes the score matching objective tractable.

  • Conditional Techniqueextends toProbability Flow ODE

    Once the score is learned, the SDE can be replaced by an equivalent deterministic ODE.

  • VAE & ELBOcontrastLeWorldModel / SIGReg

    Both learn latent representations: the VAE reconstructs observations, LeWorldModel predicts only in latent space and relies on SIGReg against collapse.