Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287642227> ?p ?o ?g. }
Showing items 1 to 59 of
59
with 100 items per page.
- W4287642227 abstract "Neural text generation (data- or text-to-text) demonstrates remarkable performance when training data is abundant which for many applications is not the case. To collect a large corpus of parallel data, heuristic rules are often used but they inevitably let noise into the data, such as phrases in the output which cannot be explained by the input. Consequently, models pick up on the noise and may hallucinate--generate fluent but unsupported text. Our contribution is a simple but powerful technique to treat such hallucinations as a controllable aspect of the generated text, without dismissing any input and without modifying the model architecture. On the WikiBio corpus (Lebret et al., 2016), a particularly noisy dataset, we demonstrate the efficacy of the technique both in an automatic and in a human evaluation." @default.
- W4287642227 created "2022-07-25" @default.
- W4287642227 creator A5037657908 @default.
- W4287642227 date "2020-10-12" @default.
- W4287642227 modified "2023-09-29" @default.
- W4287642227 title "Controlled Hallucinations: Learning to Generate Faithfully from Noisy Data" @default.
- W4287642227 doi "https://doi.org/10.48550/arxiv.2010.05873" @default.
- W4287642227 hasPublicationYear "2020" @default.
- W4287642227 type Work @default.
- W4287642227 citedByCount "0" @default.
- W4287642227 crossrefType "posted-content" @default.
- W4287642227 hasAuthorship W4287642227A5037657908 @default.
- W4287642227 hasBestOaLocation W42876422271 @default.
- W4287642227 hasConcept C111472728 @default.
- W4287642227 hasConcept C115961682 @default.
- W4287642227 hasConcept C119857082 @default.
- W4287642227 hasConcept C138885662 @default.
- W4287642227 hasConcept C153180895 @default.
- W4287642227 hasConcept C154945302 @default.
- W4287642227 hasConcept C173801870 @default.
- W4287642227 hasConcept C204321447 @default.
- W4287642227 hasConcept C2780586882 @default.
- W4287642227 hasConcept C28490314 @default.
- W4287642227 hasConcept C2911011789 @default.
- W4287642227 hasConcept C2985684807 @default.
- W4287642227 hasConcept C41008148 @default.
- W4287642227 hasConcept C51632099 @default.
- W4287642227 hasConcept C99498987 @default.
- W4287642227 hasConceptScore W4287642227C111472728 @default.
- W4287642227 hasConceptScore W4287642227C115961682 @default.
- W4287642227 hasConceptScore W4287642227C119857082 @default.
- W4287642227 hasConceptScore W4287642227C138885662 @default.
- W4287642227 hasConceptScore W4287642227C153180895 @default.
- W4287642227 hasConceptScore W4287642227C154945302 @default.
- W4287642227 hasConceptScore W4287642227C173801870 @default.
- W4287642227 hasConceptScore W4287642227C204321447 @default.
- W4287642227 hasConceptScore W4287642227C2780586882 @default.
- W4287642227 hasConceptScore W4287642227C28490314 @default.
- W4287642227 hasConceptScore W4287642227C2911011789 @default.
- W4287642227 hasConceptScore W4287642227C2985684807 @default.
- W4287642227 hasConceptScore W4287642227C41008148 @default.
- W4287642227 hasConceptScore W4287642227C51632099 @default.
- W4287642227 hasConceptScore W4287642227C99498987 @default.
- W4287642227 hasLocation W42876422271 @default.
- W4287642227 hasOpenAccess W4287642227 @default.
- W4287642227 hasPrimaryLocation W42876422271 @default.
- W4287642227 hasRelatedWork W10324545 @default.
- W4287642227 hasRelatedWork W10648265 @default.
- W4287642227 hasRelatedWork W12361542 @default.
- W4287642227 hasRelatedWork W1614943 @default.
- W4287642227 hasRelatedWork W449952 @default.
- W4287642227 hasRelatedWork W5192282 @default.
- W4287642227 hasRelatedWork W6869908 @default.
- W4287642227 hasRelatedWork W7655147 @default.
- W4287642227 hasRelatedWork W7982726 @default.
- W4287642227 hasRelatedWork W8136493 @default.
- W4287642227 isParatext "false" @default.
- W4287642227 isRetracted "false" @default.
- W4287642227 workType "article" @default.