Matches in SemOpenAlex for { <https://semopenalex.org/work/W3039449038> ?p ?o ?g. }
- W3039449038 abstract "This paper combines the advantages of parametric and non-parametric, exemplar based generative models using variational inference, yielding a new generative model called Exemplar VAE. This is a variant of VAE with a non-parametric Parzen window prior in the latent space. To sample from it, one first draws a random exemplar from training data, then stochastically transforms the exemplar into a latent code and a new observation. We propose Retrieval Augmented Training (RAT) that uses approximate nearest neighbor search in the latent space to speed up training based on a novel lower bound on log marginal likelihood. To enhance generalization, model parameters are learned using exemplar leave-one-out and subsampling. Experiments demonstrate the effectiveness of Exemplar VAEs on density estimation and representation learning. Further, generative data augmentation using Exemplar VAEs on permutation invariant MNIST and Fashion MNIST reduces classification error from 1.23 to 0.69 and 8.56 to 8.16." @default.
- W3039449038 created "2020-07-10" @default.
- W3039449038 creator A5018912210 @default.
- W3039449038 creator A5035123135 @default.
- W3039449038 creator A5046086284 @default.
- W3039449038 date "2020-04-09" @default.
- W3039449038 modified "2023-09-27" @default.
- W3039449038 title "Exemplar based Generation and Data Augmentation using Exemplar VAEs" @default.
- W3039449038 cites W1516111018 @default.
- W3039449038 cites W1583912456 @default.
- W3039449038 cites W1959608418 @default.
- W3039449038 cites W2086504823 @default.
- W3039449038 cites W2097074225 @default.
- W3039449038 cites W2116013899 @default.
- W3039449038 cites W2117154949 @default.
- W3039449038 cites W2118020555 @default.
- W3039449038 cites W2125873654 @default.
- W3039449038 cites W2147880316 @default.
- W3039449038 cites W2171011251 @default.
- W3039449038 cites W2331128040 @default.
- W3039449038 cites W2557579533 @default.
- W3039449038 cites W2560512785 @default.
- W3039449038 cites W2753738274 @default.
- W3039449038 cites W2785921341 @default.
- W3039449038 cites W2911290044 @default.
- W3039449038 cites W2949354227 @default.
- W3039449038 cites W2949899814 @default.
- W3039449038 cites W2962738009 @default.
- W3039449038 cites W2963018920 @default.
- W3039449038 cites W2963082441 @default.
- W3039449038 cites W2963207607 @default.
- W3039449038 cites W2963970792 @default.
- W3039449038 cites W2963981733 @default.
- W3039449038 cites W2964121744 @default.
- W3039449038 cites W2964122153 @default.
- W3039449038 cites W2964127395 @default.
- W3039449038 cites W2964159205 @default.
- W3039449038 cites W2990155616 @default.
- W3039449038 cites W2995154514 @default.
- W3039449038 hasPublicationYear "2020" @default.
- W3039449038 type Work @default.
- W3039449038 sameAs 3039449038 @default.
- W3039449038 citedByCount "0" @default.
- W3039449038 crossrefType "posted-content" @default.
- W3039449038 hasAuthorship W3039449038A5018912210 @default.
- W3039449038 hasAuthorship W3039449038A5035123135 @default.
- W3039449038 hasAuthorship W3039449038A5046086284 @default.
- W3039449038 hasConcept C105795698 @default.
- W3039449038 hasConcept C107673813 @default.
- W3039449038 hasConcept C117251300 @default.
- W3039449038 hasConcept C119857082 @default.
- W3039449038 hasConcept C153180895 @default.
- W3039449038 hasConcept C154945302 @default.
- W3039449038 hasConcept C167966045 @default.
- W3039449038 hasConcept C185429906 @default.
- W3039449038 hasConcept C189508267 @default.
- W3039449038 hasConcept C190502265 @default.
- W3039449038 hasConcept C2776214188 @default.
- W3039449038 hasConcept C33923547 @default.
- W3039449038 hasConcept C39890363 @default.
- W3039449038 hasConcept C41008148 @default.
- W3039449038 hasConcept C50644808 @default.
- W3039449038 hasConcept C95923904 @default.
- W3039449038 hasConceptScore W3039449038C105795698 @default.
- W3039449038 hasConceptScore W3039449038C107673813 @default.
- W3039449038 hasConceptScore W3039449038C117251300 @default.
- W3039449038 hasConceptScore W3039449038C119857082 @default.
- W3039449038 hasConceptScore W3039449038C153180895 @default.
- W3039449038 hasConceptScore W3039449038C154945302 @default.
- W3039449038 hasConceptScore W3039449038C167966045 @default.
- W3039449038 hasConceptScore W3039449038C185429906 @default.
- W3039449038 hasConceptScore W3039449038C189508267 @default.
- W3039449038 hasConceptScore W3039449038C190502265 @default.
- W3039449038 hasConceptScore W3039449038C2776214188 @default.
- W3039449038 hasConceptScore W3039449038C33923547 @default.
- W3039449038 hasConceptScore W3039449038C39890363 @default.
- W3039449038 hasConceptScore W3039449038C41008148 @default.
- W3039449038 hasConceptScore W3039449038C50644808 @default.
- W3039449038 hasConceptScore W3039449038C95923904 @default.
- W3039449038 hasLocation W30394490381 @default.
- W3039449038 hasOpenAccess W3039449038 @default.
- W3039449038 hasPrimaryLocation W30394490381 @default.
- W3039449038 hasRelatedWork W1578488390 @default.
- W3039449038 hasRelatedWork W2551604710 @default.
- W3039449038 hasRelatedWork W2599907099 @default.
- W3039449038 hasRelatedWork W2604977777 @default.
- W3039449038 hasRelatedWork W2617620476 @default.
- W3039449038 hasRelatedWork W2617780315 @default.
- W3039449038 hasRelatedWork W2730106296 @default.
- W3039449038 hasRelatedWork W2771099609 @default.
- W3039449038 hasRelatedWork W2896335160 @default.
- W3039449038 hasRelatedWork W2911463906 @default.
- W3039449038 hasRelatedWork W2922338294 @default.
- W3039449038 hasRelatedWork W2933097077 @default.
- W3039449038 hasRelatedWork W2950920935 @default.
- W3039449038 hasRelatedWork W2964031641 @default.
- W3039449038 hasRelatedWork W2973021803 @default.
- W3039449038 hasRelatedWork W3040915569 @default.
- W3039449038 hasRelatedWork W3048772196 @default.
- W3039449038 hasRelatedWork W3090198400 @default.