Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287802227> ?p ?o ?g. }
Showing items 1 to 70 of
70
with 100 items per page.
- W4287802227 abstract "Variational autoencoders (VAEs) are a standard framework for inducing latent variable models that have been shown effective in learning text representations as well as in text generation. The key challenge with using VAEs is the {it posterior collapse} problem: learning tends to converge to trivial solutions where the generators ignore latent variables. In our Levenstein VAE, we propose to replace the evidence lower bound (ELBO) with a new objective which is simple to optimize and prevents posterior collapse. Intuitively, it corresponds to generating a sequence from the autoencoder and encouraging the model to predict an optimal continuation according to the Levenshtein distance (LD) with the reference sentence at each time step in the generated sequence. We motivate the method from the probabilistic perspective by showing that it is closely related to optimizing a bound on the intractable Kullback-Leibler divergence of an LD-based kernel density estimator from the model distribution. With this objective, any generator disregarding latent variables will incur large penalties and hence posterior collapse does not happen. We relate our approach to policy distillation cite{RossGB11} and dynamic oracles cite{GoldbergN12}. By considering Yelp and SNLI benchmarks, we show that Levenstein VAE produces more informative latent representations than alternative approaches to preventing posterior collapse." @default.
- W4287802227 created "2022-07-26" @default.
- W4287802227 creator A5065341276 @default.
- W4287802227 creator A5086717154 @default.
- W4287802227 date "2020-04-30" @default.
- W4287802227 modified "2023-09-28" @default.
- W4287802227 title "Preventing Posterior Collapse with Levenshtein Variational Autoencoder" @default.
- W4287802227 doi "https://doi.org/10.48550/arxiv.2004.14758" @default.
- W4287802227 hasPublicationYear "2020" @default.
- W4287802227 type Work @default.
- W4287802227 citedByCount "1" @default.
- W4287802227 countsByYear W42878022272023 @default.
- W4287802227 crossrefType "posted-content" @default.
- W4287802227 hasAuthorship W4287802227A5065341276 @default.
- W4287802227 hasAuthorship W4287802227A5086717154 @default.
- W4287802227 hasBestOaLocation W42878022271 @default.
- W4287802227 hasConcept C101738243 @default.
- W4287802227 hasConcept C105795698 @default.
- W4287802227 hasConcept C107673813 @default.
- W4287802227 hasConcept C108583219 @default.
- W4287802227 hasConcept C119857082 @default.
- W4287802227 hasConcept C134306372 @default.
- W4287802227 hasConcept C138885662 @default.
- W4287802227 hasConcept C154945302 @default.
- W4287802227 hasConcept C185429906 @default.
- W4287802227 hasConcept C207390915 @default.
- W4287802227 hasConcept C2778112365 @default.
- W4287802227 hasConcept C33923547 @default.
- W4287802227 hasConcept C41008148 @default.
- W4287802227 hasConcept C41895202 @default.
- W4287802227 hasConcept C51167844 @default.
- W4287802227 hasConcept C54355233 @default.
- W4287802227 hasConcept C57830394 @default.
- W4287802227 hasConcept C77553402 @default.
- W4287802227 hasConcept C86803240 @default.
- W4287802227 hasConceptScore W4287802227C101738243 @default.
- W4287802227 hasConceptScore W4287802227C105795698 @default.
- W4287802227 hasConceptScore W4287802227C107673813 @default.
- W4287802227 hasConceptScore W4287802227C108583219 @default.
- W4287802227 hasConceptScore W4287802227C119857082 @default.
- W4287802227 hasConceptScore W4287802227C134306372 @default.
- W4287802227 hasConceptScore W4287802227C138885662 @default.
- W4287802227 hasConceptScore W4287802227C154945302 @default.
- W4287802227 hasConceptScore W4287802227C185429906 @default.
- W4287802227 hasConceptScore W4287802227C207390915 @default.
- W4287802227 hasConceptScore W4287802227C2778112365 @default.
- W4287802227 hasConceptScore W4287802227C33923547 @default.
- W4287802227 hasConceptScore W4287802227C41008148 @default.
- W4287802227 hasConceptScore W4287802227C41895202 @default.
- W4287802227 hasConceptScore W4287802227C51167844 @default.
- W4287802227 hasConceptScore W4287802227C54355233 @default.
- W4287802227 hasConceptScore W4287802227C57830394 @default.
- W4287802227 hasConceptScore W4287802227C77553402 @default.
- W4287802227 hasConceptScore W4287802227C86803240 @default.
- W4287802227 hasLocation W42878022271 @default.
- W4287802227 hasOpenAccess W4287802227 @default.
- W4287802227 hasPrimaryLocation W42878022271 @default.
- W4287802227 hasRelatedWork W2057965038 @default.
- W4287802227 hasRelatedWork W2891583400 @default.
- W4287802227 hasRelatedWork W2899106999 @default.
- W4287802227 hasRelatedWork W2949649223 @default.
- W4287802227 hasRelatedWork W3044458868 @default.
- W4287802227 hasRelatedWork W3111317448 @default.
- W4287802227 hasRelatedWork W3183987844 @default.
- W4287802227 hasRelatedWork W4213225422 @default.
- W4287802227 hasRelatedWork W4220810357 @default.
- W4287802227 hasRelatedWork W4225782343 @default.
- W4287802227 isParatext "false" @default.
- W4287802227 isRetracted "false" @default.
- W4287802227 workType "article" @default.