Matches in SemOpenAlex for { <https://semopenalex.org/work/W4367000349> ?p ?o ?g. }
Showing items 1 to 61 of
61
with 100 items per page.
- W4367000349 abstract "As a widely recognized approach to deep generative modeling, Variational Auto-Encoders (VAEs) still face challenges with the quality of generated images, often presenting noticeable blurriness. This issue stems from the unrealistic assumption that approximates the conditional data distribution, $p(textbf{x} | textbf{z})$, as an isotropic Gaussian. In this paper, we propose a novel solution to address these issues. We illustrate how one can extract a latent space from a pre-existing diffusion model by optimizing an encoder to maximize the marginal data log-likelihood. Furthermore, we demonstrate that a decoder can be analytically derived post encoder-training, employing the Bayes rule for scores. This leads to a VAE-esque deep latent variable model, which discards the need for Gaussian assumptions on $p(textbf{x} | textbf{z})$ or the training of a separate decoder network. Our method, which capitalizes on the strengths of pre-trained diffusion models and equips them with latent spaces, results in a significant enhancement to the performance of VAEs." @default.
- W4367000349 created "2023-04-27" @default.
- W4367000349 creator A5033880300 @default.
- W4367000349 creator A5041838383 @default.
- W4367000349 creator A5051470659 @default.
- W4367000349 date "2023-04-24" @default.
- W4367000349 modified "2023-09-29" @default.
- W4367000349 title "Variational Diffusion Auto-encoder: Latent Space Extraction from Pre-trained Diffusion Models" @default.
- W4367000349 doi "https://doi.org/10.48550/arxiv.2304.12141" @default.
- W4367000349 hasPublicationYear "2023" @default.
- W4367000349 type Work @default.
- W4367000349 citedByCount "0" @default.
- W4367000349 crossrefType "posted-content" @default.
- W4367000349 hasAuthorship W4367000349A5033880300 @default.
- W4367000349 hasAuthorship W4367000349A5041838383 @default.
- W4367000349 hasAuthorship W4367000349A5051470659 @default.
- W4367000349 hasBestOaLocation W43670003491 @default.
- W4367000349 hasConcept C111919701 @default.
- W4367000349 hasConcept C11413529 @default.
- W4367000349 hasConcept C118505674 @default.
- W4367000349 hasConcept C121332964 @default.
- W4367000349 hasConcept C153180895 @default.
- W4367000349 hasConcept C154945302 @default.
- W4367000349 hasConcept C163716315 @default.
- W4367000349 hasConcept C184050105 @default.
- W4367000349 hasConcept C2778572836 @default.
- W4367000349 hasConcept C41008148 @default.
- W4367000349 hasConcept C51167844 @default.
- W4367000349 hasConcept C62520636 @default.
- W4367000349 hasConcept C69357855 @default.
- W4367000349 hasConcept C97355855 @default.
- W4367000349 hasConceptScore W4367000349C111919701 @default.
- W4367000349 hasConceptScore W4367000349C11413529 @default.
- W4367000349 hasConceptScore W4367000349C118505674 @default.
- W4367000349 hasConceptScore W4367000349C121332964 @default.
- W4367000349 hasConceptScore W4367000349C153180895 @default.
- W4367000349 hasConceptScore W4367000349C154945302 @default.
- W4367000349 hasConceptScore W4367000349C163716315 @default.
- W4367000349 hasConceptScore W4367000349C184050105 @default.
- W4367000349 hasConceptScore W4367000349C2778572836 @default.
- W4367000349 hasConceptScore W4367000349C41008148 @default.
- W4367000349 hasConceptScore W4367000349C51167844 @default.
- W4367000349 hasConceptScore W4367000349C62520636 @default.
- W4367000349 hasConceptScore W4367000349C69357855 @default.
- W4367000349 hasConceptScore W4367000349C97355855 @default.
- W4367000349 hasLocation W43670003491 @default.
- W4367000349 hasOpenAccess W4367000349 @default.
- W4367000349 hasPrimaryLocation W43670003491 @default.
- W4367000349 hasRelatedWork W2033914206 @default.
- W4367000349 hasRelatedWork W2053555031 @default.
- W4367000349 hasRelatedWork W2356875448 @default.
- W4367000349 hasRelatedWork W2358489738 @default.
- W4367000349 hasRelatedWork W2385621972 @default.
- W4367000349 hasRelatedWork W2389148534 @default.
- W4367000349 hasRelatedWork W2753215597 @default.
- W4367000349 hasRelatedWork W2963594498 @default.
- W4367000349 hasRelatedWork W3107474891 @default.
- W4367000349 hasRelatedWork W3126277921 @default.
- W4367000349 isParatext "false" @default.
- W4367000349 isRetracted "false" @default.
- W4367000349 workType "article" @default.