Matches in SemOpenAlex for { <https://semopenalex.org/work/W3118332005> ?p ?o ?g. }
- W3118332005 endingPage "593" @default.
- W3118332005 startingPage "584" @default.
- W3118332005 abstract "Pre-training models on vast quantities of unlabeled data has emerged as an effective approach to improving accuracy on many NLP tasks. On the other hand, traditional machine translation has a long history of leveraging unlabeled data through noisy channel modeling. The same idea has recently been shown to achieve strong improvements for neural machine translation. Unfortunately, na ive noisy channel modeling with modern sequence to sequence models is up to an order of magnitude slower than alternatives. We address this issue by introducing efficient approximations to make inference with the noisy channel approach as fast as strong ensembles while increasing accuracy. We also show that the noisy channel approach can outperform strong pre-training results by achieving a new state of the art on WMT Romanian-English translation." @default.
- W3118332005 created "2021-01-18" @default.
- W3118332005 creator A5016113002 @default.
- W3118332005 creator A5052127592 @default.
- W3118332005 creator A5065321401 @default.
- W3118332005 creator A5083212922 @default.
- W3118332005 date "2020-11-01" @default.
- W3118332005 modified "2023-09-30" @default.
- W3118332005 title "Language Models not just for Pre-training: Fast Online Neural Noisy Channel Modeling" @default.
- W3118332005 cites W1915251500 @default.
- W3118332005 cites W2006969979 @default.
- W3118332005 cites W211509693 @default.
- W3118332005 cites W2124807415 @default.
- W3118332005 cites W2153653739 @default.
- W3118332005 cites W2889326796 @default.
- W3118332005 cites W2908336025 @default.
- W3118332005 cites W2922709902 @default.
- W3118332005 cites W2933138175 @default.
- W3118332005 cites W2962784628 @default.
- W3118332005 cites W2962801832 @default.
- W3118332005 cites W2963174344 @default.
- W3118332005 cites W2963216553 @default.
- W3118332005 cites W2963341956 @default.
- W3118332005 cites W2963403868 @default.
- W3118332005 cites W2963532001 @default.
- W3118332005 cites W2963807318 @default.
- W3118332005 cites W2963833497 @default.
- W3118332005 cites W2963937700 @default.
- W3118332005 cites W2970295111 @default.
- W3118332005 cites W2970694516 @default.
- W3118332005 cites W2983040767 @default.
- W3118332005 cites W2988615798 @default.
- W3118332005 cites W2988739750 @default.
- W3118332005 cites W2995746049 @default.
- W3118332005 cites W2995983533 @default.
- W3118332005 cites W2996159613 @default.
- W3118332005 cites W3032816972 @default.
- W3118332005 cites W3035812575 @default.
- W3118332005 cites W3042199843 @default.
- W3118332005 cites W3082274269 @default.
- W3118332005 cites W3107826490 @default.
- W3118332005 hasPublicationYear "2020" @default.
- W3118332005 type Work @default.
- W3118332005 sameAs 3118332005 @default.
- W3118332005 citedByCount "0" @default.
- W3118332005 crossrefType "proceedings-article" @default.
- W3118332005 hasAuthorship W3118332005A5016113002 @default.
- W3118332005 hasAuthorship W3118332005A5052127592 @default.
- W3118332005 hasAuthorship W3118332005A5065321401 @default.
- W3118332005 hasAuthorship W3118332005A5083212922 @default.
- W3118332005 hasConcept C104317684 @default.
- W3118332005 hasConcept C105580179 @default.
- W3118332005 hasConcept C119857082 @default.
- W3118332005 hasConcept C121332964 @default.
- W3118332005 hasConcept C127162648 @default.
- W3118332005 hasConcept C137293760 @default.
- W3118332005 hasConcept C149364088 @default.
- W3118332005 hasConcept C153294291 @default.
- W3118332005 hasConcept C154945302 @default.
- W3118332005 hasConcept C185592680 @default.
- W3118332005 hasConcept C203005215 @default.
- W3118332005 hasConcept C204321447 @default.
- W3118332005 hasConcept C2776214188 @default.
- W3118332005 hasConcept C2777211547 @default.
- W3118332005 hasConcept C2778112365 @default.
- W3118332005 hasConcept C2781170535 @default.
- W3118332005 hasConcept C31258907 @default.
- W3118332005 hasConcept C41008148 @default.
- W3118332005 hasConcept C51632099 @default.
- W3118332005 hasConcept C54355233 @default.
- W3118332005 hasConcept C55493867 @default.
- W3118332005 hasConcept C86803240 @default.
- W3118332005 hasConceptScore W3118332005C104317684 @default.
- W3118332005 hasConceptScore W3118332005C105580179 @default.
- W3118332005 hasConceptScore W3118332005C119857082 @default.
- W3118332005 hasConceptScore W3118332005C121332964 @default.
- W3118332005 hasConceptScore W3118332005C127162648 @default.
- W3118332005 hasConceptScore W3118332005C137293760 @default.
- W3118332005 hasConceptScore W3118332005C149364088 @default.
- W3118332005 hasConceptScore W3118332005C153294291 @default.
- W3118332005 hasConceptScore W3118332005C154945302 @default.
- W3118332005 hasConceptScore W3118332005C185592680 @default.
- W3118332005 hasConceptScore W3118332005C203005215 @default.
- W3118332005 hasConceptScore W3118332005C204321447 @default.
- W3118332005 hasConceptScore W3118332005C2776214188 @default.
- W3118332005 hasConceptScore W3118332005C2777211547 @default.
- W3118332005 hasConceptScore W3118332005C2778112365 @default.
- W3118332005 hasConceptScore W3118332005C2781170535 @default.
- W3118332005 hasConceptScore W3118332005C31258907 @default.
- W3118332005 hasConceptScore W3118332005C41008148 @default.
- W3118332005 hasConceptScore W3118332005C51632099 @default.
- W3118332005 hasConceptScore W3118332005C54355233 @default.
- W3118332005 hasConceptScore W3118332005C55493867 @default.
- W3118332005 hasConceptScore W3118332005C86803240 @default.
- W3118332005 hasLocation W31183320051 @default.
- W3118332005 hasOpenAccess W3118332005 @default.
- W3118332005 hasPrimaryLocation W31183320051 @default.
- W3118332005 hasRelatedWork W1866072925 @default.