Matches in SemOpenAlex for { <https://semopenalex.org/work/W3030343076> ?p ?o ?g. }
- W3030343076 abstract "The two dominant approaches to neural text generation are fully autoregressive models, using serial beam search decoding, and non-autoregressive models, using parallel decoding with no output dependencies. This work proposes an autoregressive model with sub-linear parallel time generation. Noting that conditional random fields with bounded context can be decoded in parallel, we propose an efficient cascaded decoding approach for generating high-quality output. To parameterize this cascade, we introduce a Markov transformer, a variant of the popular fully autoregressive model that allows us to simultaneously decode with specific autoregressive context cutoffs. This approach requires only a small modification from standard autoregressive training, while showing competitive accuracy/speed tradeoff compared to existing methods on five machine translation datasets." @default.
- W3030343076 created "2020-06-05" @default.
- W3030343076 creator A5085355324 @default.
- W3030343076 creator A5088888805 @default.
- W3030343076 date "2020-06-01" @default.
- W3030343076 modified "2023-09-23" @default.
- W3030343076 title "Cascaded Text Generation with Markov Transformers" @default.
- W3030343076 cites W1514535095 @default.
- W3030343076 cites W1522301498 @default.
- W3030343076 cites W1816313093 @default.
- W3030343076 cites W1821462560 @default.
- W3030343076 cites W1843891098 @default.
- W3030343076 cites W1902237438 @default.
- W3030343076 cites W1959608418 @default.
- W3030343076 cites W2064675550 @default.
- W3030343076 cites W2101714644 @default.
- W3030343076 cites W2119052294 @default.
- W3030343076 cites W2122262818 @default.
- W3030343076 cites W2125712079 @default.
- W3030343076 cites W2133564696 @default.
- W3030343076 cites W2147880316 @default.
- W3030343076 cites W2152966212 @default.
- W3030343076 cites W2178449955 @default.
- W3030343076 cites W2251610689 @default.
- W3030343076 cites W2402268235 @default.
- W3030343076 cites W2463507112 @default.
- W3030343076 cites W2508316494 @default.
- W3030343076 cites W2514713644 @default.
- W3030343076 cites W2606974598 @default.
- W3030343076 cites W2767206889 @default.
- W3030343076 cites W2885185669 @default.
- W3030343076 cites W2890438841 @default.
- W3030343076 cites W2920538220 @default.
- W3030343076 cites W2937808806 @default.
- W3030343076 cites W2945928904 @default.
- W3030343076 cites W2946375144 @default.
- W3030343076 cites W2947334393 @default.
- W3030343076 cites W2947898088 @default.
- W3030343076 cites W2948381505 @default.
- W3030343076 cites W2948975009 @default.
- W3030343076 cites W2949417144 @default.
- W3030343076 cites W2951080837 @default.
- W3030343076 cites W2951655021 @default.
- W3030343076 cites W2952509486 @default.
- W3030343076 cites W2962897886 @default.
- W3030343076 cites W2962969034 @default.
- W3030343076 cites W2963403868 @default.
- W3030343076 cites W2963434219 @default.
- W3030343076 cites W2963536265 @default.
- W3030343076 cites W2963946353 @default.
- W3030343076 cites W2968297680 @default.
- W3030343076 cites W2969263964 @default.
- W3030343076 cites W2969958763 @default.
- W3030343076 cites W2970206392 @default.
- W3030343076 cites W2970690146 @default.
- W3030343076 cites W2970971581 @default.
- W3030343076 cites W2971008823 @default.
- W3030343076 cites W2972034672 @default.
- W3030343076 cites W2973727699 @default.
- W3030343076 cites W2985694911 @default.
- W3030343076 cites W2986772082 @default.
- W3030343076 cites W2990389671 @default.
- W3030343076 cites W3000840023 @default.
- W3030343076 cites W3008170323 @default.
- W3030343076 cites W3008388412 @default.
- W3030343076 cites W3021914396 @default.
- W3030343076 cites W3022593918 @default.
- W3030343076 cites W3030163527 @default.
- W3030343076 cites W3041866211 @default.
- W3030343076 cites W3087983746 @default.
- W3030343076 cites W3100753857 @default.
- W3030343076 cites W3106234277 @default.
- W3030343076 cites W3141754740 @default.
- W3030343076 cites W3004162225 @default.
- W3030343076 hasPublicationYear "2020" @default.
- W3030343076 type Work @default.
- W3030343076 sameAs 3030343076 @default.
- W3030343076 citedByCount "6" @default.
- W3030343076 countsByYear W30303430762020 @default.
- W3030343076 countsByYear W30303430762021 @default.
- W3030343076 crossrefType "posted-content" @default.
- W3030343076 hasAuthorship W3030343076A5085355324 @default.
- W3030343076 hasAuthorship W3030343076A5088888805 @default.
- W3030343076 hasConcept C11413529 @default.
- W3030343076 hasConcept C119857082 @default.
- W3030343076 hasConcept C121332964 @default.
- W3030343076 hasConcept C149782125 @default.
- W3030343076 hasConcept C151406439 @default.
- W3030343076 hasConcept C151730666 @default.
- W3030343076 hasConcept C154945302 @default.
- W3030343076 hasConcept C159877910 @default.
- W3030343076 hasConcept C165801399 @default.
- W3030343076 hasConcept C185592680 @default.
- W3030343076 hasConcept C194657046 @default.
- W3030343076 hasConcept C23224414 @default.
- W3030343076 hasConcept C24338571 @default.
- W3030343076 hasConcept C2779343474 @default.
- W3030343076 hasConcept C28490314 @default.
- W3030343076 hasConcept C33923547 @default.
- W3030343076 hasConcept C34146451 @default.