Matches in SemOpenAlex for { <https://semopenalex.org/work/W4288334893> ?p ?o ?g. }
Showing items 1 to 75 of
75
with 100 items per page.
- W4288334893 abstract "Energy-based models (EBMs), a.k.a. un-normalized models, have had recent successes in continuous spaces. However, they have not been successfully applied to model text sequences. While decreasing the energy at training samples is straightforward, mining (negative) samples where the energy should be increased is difficult. In part, this is because standard gradient-based methods are not readily applicable when the input is high-dimensional and discrete. Here, we side-step this issue by generating negatives using pre-trained auto-regressive language models. The EBM then works in the residual of the language model; and is trained to discriminate real text from text generated by the auto-regressive models. We investigate the generalization ability of residual EBMs, a pre-requisite for using them in other applications. We extensively analyze generalization for the task of classifying whether an input is machine or human generated, a natural task given the training loss and how we mine negatives. Overall, we observe that EBMs can generalize remarkably well to changes in the architecture of the generators producing negatives. However, EBMs exhibit more sensitivity to the training set used by such generators." @default.
- W4288334893 created "2022-07-28" @default.
- W4288334893 creator A5024680544 @default.
- W4288334893 creator A5053145694 @default.
- W4288334893 creator A5076248976 @default.
- W4288334893 creator A5077887756 @default.
- W4288334893 creator A5084196073 @default.
- W4288334893 creator A5088888805 @default.
- W4288334893 date "2019-06-07" @default.
- W4288334893 modified "2023-09-27" @default.
- W4288334893 title "Real or Fake? Learning to Discriminate Machine from Human Generated Text" @default.
- W4288334893 doi "https://doi.org/10.48550/arxiv.1906.03351" @default.
- W4288334893 hasPublicationYear "2019" @default.
- W4288334893 type Work @default.
- W4288334893 citedByCount "0" @default.
- W4288334893 crossrefType "posted-content" @default.
- W4288334893 hasAuthorship W4288334893A5024680544 @default.
- W4288334893 hasAuthorship W4288334893A5053145694 @default.
- W4288334893 hasAuthorship W4288334893A5076248976 @default.
- W4288334893 hasAuthorship W4288334893A5077887756 @default.
- W4288334893 hasAuthorship W4288334893A5084196073 @default.
- W4288334893 hasAuthorship W4288334893A5088888805 @default.
- W4288334893 hasBestOaLocation W42883348931 @default.
- W4288334893 hasConcept C105795698 @default.
- W4288334893 hasConcept C11413529 @default.
- W4288334893 hasConcept C119857082 @default.
- W4288334893 hasConcept C127413603 @default.
- W4288334893 hasConcept C134306372 @default.
- W4288334893 hasConcept C153180895 @default.
- W4288334893 hasConcept C154945302 @default.
- W4288334893 hasConcept C155512373 @default.
- W4288334893 hasConcept C177148314 @default.
- W4288334893 hasConcept C177264268 @default.
- W4288334893 hasConcept C186370098 @default.
- W4288334893 hasConcept C199360897 @default.
- W4288334893 hasConcept C201995342 @default.
- W4288334893 hasConcept C204321447 @default.
- W4288334893 hasConcept C2780451532 @default.
- W4288334893 hasConcept C33923547 @default.
- W4288334893 hasConcept C41008148 @default.
- W4288334893 hasConcept C51632099 @default.
- W4288334893 hasConceptScore W4288334893C105795698 @default.
- W4288334893 hasConceptScore W4288334893C11413529 @default.
- W4288334893 hasConceptScore W4288334893C119857082 @default.
- W4288334893 hasConceptScore W4288334893C127413603 @default.
- W4288334893 hasConceptScore W4288334893C134306372 @default.
- W4288334893 hasConceptScore W4288334893C153180895 @default.
- W4288334893 hasConceptScore W4288334893C154945302 @default.
- W4288334893 hasConceptScore W4288334893C155512373 @default.
- W4288334893 hasConceptScore W4288334893C177148314 @default.
- W4288334893 hasConceptScore W4288334893C177264268 @default.
- W4288334893 hasConceptScore W4288334893C186370098 @default.
- W4288334893 hasConceptScore W4288334893C199360897 @default.
- W4288334893 hasConceptScore W4288334893C201995342 @default.
- W4288334893 hasConceptScore W4288334893C204321447 @default.
- W4288334893 hasConceptScore W4288334893C2780451532 @default.
- W4288334893 hasConceptScore W4288334893C33923547 @default.
- W4288334893 hasConceptScore W4288334893C41008148 @default.
- W4288334893 hasConceptScore W4288334893C51632099 @default.
- W4288334893 hasLocation W42883348931 @default.
- W4288334893 hasOpenAccess W4288334893 @default.
- W4288334893 hasPrimaryLocation W42883348931 @default.
- W4288334893 hasRelatedWork W1509467138 @default.
- W4288334893 hasRelatedWork W1645673120 @default.
- W4288334893 hasRelatedWork W2081647779 @default.
- W4288334893 hasRelatedWork W2346074333 @default.
- W4288334893 hasRelatedWork W2989932438 @default.
- W4288334893 hasRelatedWork W3090337104 @default.
- W4288334893 hasRelatedWork W3185852197 @default.
- W4288334893 hasRelatedWork W3193088696 @default.
- W4288334893 hasRelatedWork W3201070945 @default.
- W4288334893 hasRelatedWork W4301184459 @default.
- W4288334893 isParatext "false" @default.
- W4288334893 isRetracted "false" @default.
- W4288334893 workType "article" @default.