Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385569892> ?p ?o ?g. }
Showing items 1 to 79 of
79
with 100 items per page.
- W4385569892 abstract "Due to the huge amount of parameters, finetuning of pretrained language models (PLMs) is prone to overfitting in the low resource scenarios. In this work, we present a novel method that operates on the hidden representations of a PLM to reduce overfitting. During fine-tuning, our method inserts random autoencoders between the hidden layers of a PLM, which transform activations from the previous layers into multi-view compressed representations before feeding them into the upper layers. The autoencoders are plugged out after fine-tuning, so our method does not add extra parameters or increase computation cost during inference. Our method demonstrates promising performance improvement across a wide range of sequence- and token-level lowresource NLP tasks." @default.
- W4385569892 created "2023-08-05" @default.
- W4385569892 creator A5005443526 @default.
- W4385569892 creator A5022526821 @default.
- W4385569892 creator A5045032553 @default.
- W4385569892 creator A5046532741 @default.
- W4385569892 creator A5046872618 @default.
- W4385569892 creator A5057035600 @default.
- W4385569892 creator A5086674741 @default.
- W4385569892 date "2023-01-01" @default.
- W4385569892 modified "2023-10-18" @default.
- W4385569892 title "Towards Robust Low-Resource Fine-Tuning with Multi-View Compressed Representations" @default.
- W4385569892 doi "https://doi.org/10.18653/v1/2023.acl-long.264" @default.
- W4385569892 hasPublicationYear "2023" @default.
- W4385569892 type Work @default.
- W4385569892 citedByCount "0" @default.
- W4385569892 crossrefType "proceedings-article" @default.
- W4385569892 hasAuthorship W4385569892A5005443526 @default.
- W4385569892 hasAuthorship W4385569892A5022526821 @default.
- W4385569892 hasAuthorship W4385569892A5045032553 @default.
- W4385569892 hasAuthorship W4385569892A5046532741 @default.
- W4385569892 hasAuthorship W4385569892A5046872618 @default.
- W4385569892 hasAuthorship W4385569892A5057035600 @default.
- W4385569892 hasAuthorship W4385569892A5086674741 @default.
- W4385569892 hasBestOaLocation W43855698921 @default.
- W4385569892 hasConcept C11413529 @default.
- W4385569892 hasConcept C119857082 @default.
- W4385569892 hasConcept C121332964 @default.
- W4385569892 hasConcept C127413603 @default.
- W4385569892 hasConcept C146978453 @default.
- W4385569892 hasConcept C153180895 @default.
- W4385569892 hasConcept C154945302 @default.
- W4385569892 hasConcept C157524613 @default.
- W4385569892 hasConcept C204323151 @default.
- W4385569892 hasConcept C206345919 @default.
- W4385569892 hasConcept C22019652 @default.
- W4385569892 hasConcept C2776214188 @default.
- W4385569892 hasConcept C31258907 @default.
- W4385569892 hasConcept C38652104 @default.
- W4385569892 hasConcept C41008148 @default.
- W4385569892 hasConcept C45374587 @default.
- W4385569892 hasConcept C48145219 @default.
- W4385569892 hasConcept C50644808 @default.
- W4385569892 hasConcept C62520636 @default.
- W4385569892 hasConceptScore W4385569892C11413529 @default.
- W4385569892 hasConceptScore W4385569892C119857082 @default.
- W4385569892 hasConceptScore W4385569892C121332964 @default.
- W4385569892 hasConceptScore W4385569892C127413603 @default.
- W4385569892 hasConceptScore W4385569892C146978453 @default.
- W4385569892 hasConceptScore W4385569892C153180895 @default.
- W4385569892 hasConceptScore W4385569892C154945302 @default.
- W4385569892 hasConceptScore W4385569892C157524613 @default.
- W4385569892 hasConceptScore W4385569892C204323151 @default.
- W4385569892 hasConceptScore W4385569892C206345919 @default.
- W4385569892 hasConceptScore W4385569892C22019652 @default.
- W4385569892 hasConceptScore W4385569892C2776214188 @default.
- W4385569892 hasConceptScore W4385569892C31258907 @default.
- W4385569892 hasConceptScore W4385569892C38652104 @default.
- W4385569892 hasConceptScore W4385569892C41008148 @default.
- W4385569892 hasConceptScore W4385569892C45374587 @default.
- W4385569892 hasConceptScore W4385569892C48145219 @default.
- W4385569892 hasConceptScore W4385569892C50644808 @default.
- W4385569892 hasConceptScore W4385569892C62520636 @default.
- W4385569892 hasLocation W43855698921 @default.
- W4385569892 hasOpenAccess W4385569892 @default.
- W4385569892 hasPrimaryLocation W43855698921 @default.
- W4385569892 hasRelatedWork W1996541855 @default.
- W4385569892 hasRelatedWork W2753840555 @default.
- W4385569892 hasRelatedWork W2767651786 @default.
- W4385569892 hasRelatedWork W2795435272 @default.
- W4385569892 hasRelatedWork W2940336242 @default.
- W4385569892 hasRelatedWork W2951851447 @default.
- W4385569892 hasRelatedWork W2989932438 @default.
- W4385569892 hasRelatedWork W3099765033 @default.
- W4385569892 hasRelatedWork W4210794429 @default.
- W4385569892 hasRelatedWork W4319996782 @default.
- W4385569892 isParatext "false" @default.
- W4385569892 isRetracted "false" @default.
- W4385569892 workType "article" @default.