Matches in SemOpenAlex for { <https://semopenalex.org/work/W4323651639> ?p ?o ?g. }
Showing items 1 to 57 of
57
with 100 items per page.
- W4323651639 abstract "Deep neural networks can learn powerful prior probability models for images, as evidenced by the high-quality generations obtained with recent score-based diffusion methods. But the means by which these networks capture complex global statistical structure, apparently without suffering from the curse of dimensionality, remain a mystery. To study this, we incorporate diffusion methods into a multi-scale decomposition, reducing dimensionality by assuming a stationary local Markov model for wavelet coefficients conditioned on coarser-scale coefficients. We instantiate this model using convolutional neural networks (CNNs) with local receptive fields, which enforce both the stationarity and Markov properties. Global structures are captured using a CNN with receptive fields covering the entire (but small) low-pass image. We test this model on a dataset of face images, which are highly non-stationary and contain large-scale geometric structures. Remarkably, denoising, super-resolution, and image synthesis results all demonstrate that these structures can be captured with significantly smaller conditioning neighborhoods than required by a Markov model implemented in the pixel domain. Our results show that score estimation for large complex images can be reduced to low-dimensional Markov conditional models across scales, alleviating the curse of dimensionality." @default.
- W4323651639 created "2023-03-10" @default.
- W4323651639 creator A5022370843 @default.
- W4323651639 creator A5056405277 @default.
- W4323651639 creator A5065086376 @default.
- W4323651639 creator A5080476548 @default.
- W4323651639 date "2023-03-06" @default.
- W4323651639 modified "2023-10-17" @default.
- W4323651639 title "Learning multi-scale local conditional probability models of images" @default.
- W4323651639 doi "https://doi.org/10.48550/arxiv.2303.02984" @default.
- W4323651639 hasPublicationYear "2023" @default.
- W4323651639 type Work @default.
- W4323651639 citedByCount "0" @default.
- W4323651639 crossrefType "posted-content" @default.
- W4323651639 hasAuthorship W4323651639A5022370843 @default.
- W4323651639 hasAuthorship W4323651639A5056405277 @default.
- W4323651639 hasAuthorship W4323651639A5065086376 @default.
- W4323651639 hasAuthorship W4323651639A5080476548 @default.
- W4323651639 hasBestOaLocation W43236516391 @default.
- W4323651639 hasConcept C111030470 @default.
- W4323651639 hasConcept C11413529 @default.
- W4323651639 hasConcept C119857082 @default.
- W4323651639 hasConcept C153180895 @default.
- W4323651639 hasConcept C154945302 @default.
- W4323651639 hasConcept C23224414 @default.
- W4323651639 hasConcept C33923547 @default.
- W4323651639 hasConcept C41008148 @default.
- W4323651639 hasConcept C70518039 @default.
- W4323651639 hasConcept C81363708 @default.
- W4323651639 hasConcept C98763669 @default.
- W4323651639 hasConceptScore W4323651639C111030470 @default.
- W4323651639 hasConceptScore W4323651639C11413529 @default.
- W4323651639 hasConceptScore W4323651639C119857082 @default.
- W4323651639 hasConceptScore W4323651639C153180895 @default.
- W4323651639 hasConceptScore W4323651639C154945302 @default.
- W4323651639 hasConceptScore W4323651639C23224414 @default.
- W4323651639 hasConceptScore W4323651639C33923547 @default.
- W4323651639 hasConceptScore W4323651639C41008148 @default.
- W4323651639 hasConceptScore W4323651639C70518039 @default.
- W4323651639 hasConceptScore W4323651639C81363708 @default.
- W4323651639 hasConceptScore W4323651639C98763669 @default.
- W4323651639 hasLocation W43236516391 @default.
- W4323651639 hasOpenAccess W4323651639 @default.
- W4323651639 hasPrimaryLocation W43236516391 @default.
- W4323651639 hasRelatedWork W2095834362 @default.
- W4323651639 hasRelatedWork W2110350792 @default.
- W4323651639 hasRelatedWork W2175746458 @default.
- W4323651639 hasRelatedWork W2406522397 @default.
- W4323651639 hasRelatedWork W2510756328 @default.
- W4323651639 hasRelatedWork W2613736958 @default.
- W4323651639 hasRelatedWork W2732542196 @default.
- W4323651639 hasRelatedWork W2781623059 @default.
- W4323651639 hasRelatedWork W3093612317 @default.
- W4323651639 hasRelatedWork W3211035526 @default.
- W4323651639 isParatext "false" @default.
- W4323651639 isRetracted "false" @default.
- W4323651639 workType "article" @default.