Matches in SemOpenAlex for { <https://semopenalex.org/work/W2971304722> ?p ?o ?g. }
Showing items 1 to 98 of
98
with 100 items per page.
- W2971304722 abstract "Learning target side syntactic structure has been shown to improve Neural Machine Translation (NMT). However, incorporating syntax through latent variables introduces additional complexity in inference, as the models need to marginalize over the latent syntactic structures. To avoid this, models often resort to greedy search which only allows them to explore a limited portion of the latent space. In this work, we introduce a new latent variable model, LaSyn, that captures the co-dependence between syntax and semantics, while allowing for effective and efficient inference over the latent space. LaSyn decouples direct dependence between successive latent variables, which allows its decoder to exhaustively search through the latent syntactic choices, while keeping decoding speed proportional to the size of the latent variable vocabulary. We implement LaSyn by modifying a transformer-based NMT system and design a neural expectation maximization algorithm that we regularize with part-of-speech information as the latent sequences. Evaluations on four different MT tasks show that incorporating target side syntax with LaSyn improves both translation quality, and also provides an opportunity to improve diversity." @default.
- W2971304722 created "2019-09-05" @default.
- W2971304722 creator A5006822602 @default.
- W2971304722 creator A5018054686 @default.
- W2971304722 creator A5036705554 @default.
- W2971304722 creator A5054023368 @default.
- W2971304722 creator A5054069581 @default.
- W2971304722 date "2019-08-30" @default.
- W2971304722 modified "2023-10-18" @default.
- W2971304722 title "Latent Part-of-Speech Sequences for Neural Machine Translation" @default.
- W2971304722 cites W1663973292 @default.
- W2971304722 cites W1902237438 @default.
- W2971304722 cites W2053742104 @default.
- W2971304722 cites W2101105183 @default.
- W2971304722 cites W2106751371 @default.
- W2971304722 cites W2140343992 @default.
- W2971304722 cites W2166905217 @default.
- W2971304722 cites W2222235228 @default.
- W2971304722 cites W2250861254 @default.
- W2971304722 cites W2581624817 @default.
- W2971304722 cites W2594047108 @default.
- W2971304722 cites W2739894144 @default.
- W2971304722 cites W2785093437 @default.
- W2971304722 cites W2788277448 @default.
- W2971304722 cites W2798465082 @default.
- W2971304722 cites W2800490029 @default.
- W2971304722 cites W2803369080 @default.
- W2971304722 cites W2886026946 @default.
- W2971304722 cites W2889404673 @default.
- W2971304722 cites W2890908793 @default.
- W2971304722 cites W2890909908 @default.
- W2971304722 cites W2890964657 @default.
- W2971304722 cites W2902441267 @default.
- W2971304722 cites W2907222274 @default.
- W2971304722 cites W2962717763 @default.
- W2971304722 cites W2962784628 @default.
- W2971304722 cites W2962811598 @default.
- W2971304722 cites W2963011474 @default.
- W2971304722 cites W2963206148 @default.
- W2971304722 cites W2963216553 @default.
- W2971304722 cites W2963248296 @default.
- W2971304722 cites W2963403868 @default.
- W2971304722 cites W2963648186 @default.
- W2971304722 cites W2963653811 @default.
- W2971304722 cites W2963661253 @default.
- W2971304722 cites W2963876447 @default.
- W2971304722 cites W2964048171 @default.
- W2971304722 cites W2964121744 @default.
- W2971304722 cites W2964265128 @default.
- W2971304722 cites W2964308564 @default.
- W2971304722 cites W2964352247 @default.
- W2971304722 doi "https://doi.org/10.48550/arxiv.1908.11782" @default.
- W2971304722 hasPublicationYear "2019" @default.
- W2971304722 type Work @default.
- W2971304722 sameAs 2971304722 @default.
- W2971304722 citedByCount "0" @default.
- W2971304722 crossrefType "posted-content" @default.
- W2971304722 hasAuthorship W2971304722A5006822602 @default.
- W2971304722 hasAuthorship W2971304722A5018054686 @default.
- W2971304722 hasAuthorship W2971304722A5036705554 @default.
- W2971304722 hasAuthorship W2971304722A5054023368 @default.
- W2971304722 hasAuthorship W2971304722A5054069581 @default.
- W2971304722 hasBestOaLocation W29713047221 @default.
- W2971304722 hasConcept C112933361 @default.
- W2971304722 hasConcept C119857082 @default.
- W2971304722 hasConcept C154945302 @default.
- W2971304722 hasConcept C203005215 @default.
- W2971304722 hasConcept C204321447 @default.
- W2971304722 hasConcept C2776214188 @default.
- W2971304722 hasConcept C41008148 @default.
- W2971304722 hasConcept C51167844 @default.
- W2971304722 hasConcept C60048249 @default.
- W2971304722 hasConceptScore W2971304722C112933361 @default.
- W2971304722 hasConceptScore W2971304722C119857082 @default.
- W2971304722 hasConceptScore W2971304722C154945302 @default.
- W2971304722 hasConceptScore W2971304722C203005215 @default.
- W2971304722 hasConceptScore W2971304722C204321447 @default.
- W2971304722 hasConceptScore W2971304722C2776214188 @default.
- W2971304722 hasConceptScore W2971304722C41008148 @default.
- W2971304722 hasConceptScore W2971304722C51167844 @default.
- W2971304722 hasConceptScore W2971304722C60048249 @default.
- W2971304722 hasLocation W29713047221 @default.
- W2971304722 hasOpenAccess W2971304722 @default.
- W2971304722 hasPrimaryLocation W29713047221 @default.
- W2971304722 hasRelatedWork W1585034923 @default.
- W2971304722 hasRelatedWork W1603678606 @default.
- W2971304722 hasRelatedWork W2073710650 @default.
- W2971304722 hasRelatedWork W2131933497 @default.
- W2971304722 hasRelatedWork W2161155206 @default.
- W2971304722 hasRelatedWork W2339975971 @default.
- W2971304722 hasRelatedWork W2971304722 @default.
- W2971304722 hasRelatedWork W3010693052 @default.
- W2971304722 hasRelatedWork W3107474891 @default.
- W2971304722 hasRelatedWork W2610669109 @default.
- W2971304722 isParatext "false" @default.
- W2971304722 isRetracted "false" @default.
- W2971304722 magId "2971304722" @default.
- W2971304722 workType "article" @default.