Matches in SemOpenAlex for { <https://semopenalex.org/work/W3212788173> ?p ?o ?g. }
Showing items 1 to 99 of
99
with 100 items per page.
- W3212788173 endingPage "189" @default.
- W3212788173 startingPage "181" @default.
- W3212788173 abstract "Model compression is a promising approach for reducing the model size of pretrained-language-models (PLMs) on low resource edge devices and applications. Unfortunately, the compression process always accompanies a cost of performance degradation, especially for the low resource downstream tasks, i.e., multiple-choice question answering. To address the degradation issue of model compression on PLMs, we proposed an end-to-end reptile (ETER) meta-learning approach to improving the performance of PLMs on the low resource multiple-choice question answering task. Specifically, our ETER improves the traditional two-stage meta-learning to an end-to-end manner, integrating the target finetuning stage into the meta training stage. To strengthen the generic meta-learning, ETER employs two-level meta-task construction from instance-level and domain-level to enrich its task generalization. What is more, ETER optimizes meta-learning by parameter constraints to reduce its parameter learning space. Experiments demonstrate that ETER significantly improved the performance of compressed PLMs and achieved large superiority over the baselines on different datasets." @default.
- W3212788173 created "2021-11-22" @default.
- W3212788173 creator A5009977767 @default.
- W3212788173 creator A5020987729 @default.
- W3212788173 date "2022-05-01" @default.
- W3212788173 modified "2023-10-13" @default.
- W3212788173 title "Meta-learning for compressed language model: A multiple choice question answering study" @default.
- W3212788173 cites W2091118421 @default.
- W3212788173 cites W2557764419 @default.
- W3212788173 cites W2806055002 @default.
- W3212788173 cites W2912904516 @default.
- W3212788173 cites W2949973181 @default.
- W3212788173 cites W2962925243 @default.
- W3212788173 cites W2963341956 @default.
- W3212788173 cites W2964222271 @default.
- W3212788173 cites W2997561853 @default.
- W3212788173 cites W2998183051 @default.
- W3212788173 cites W3034457371 @default.
- W3212788173 cites W3038012435 @default.
- W3212788173 cites W3099655892 @default.
- W3212788173 cites W3104216863 @default.
- W3212788173 cites W3105966348 @default.
- W3212788173 cites W3106504817 @default.
- W3212788173 cites W3177265267 @default.
- W3212788173 cites W3186799149 @default.
- W3212788173 cites W3203309275 @default.
- W3212788173 doi "https://doi.org/10.1016/j.neucom.2021.01.148" @default.
- W3212788173 hasPublicationYear "2022" @default.
- W3212788173 type Work @default.
- W3212788173 sameAs 3212788173 @default.
- W3212788173 citedByCount "3" @default.
- W3212788173 countsByYear W32127881732023 @default.
- W3212788173 crossrefType "journal-article" @default.
- W3212788173 hasAuthorship W3212788173A5009977767 @default.
- W3212788173 hasAuthorship W3212788173A5020987729 @default.
- W3212788173 hasConcept C111919701 @default.
- W3212788173 hasConcept C119857082 @default.
- W3212788173 hasConcept C134306372 @default.
- W3212788173 hasConcept C137293760 @default.
- W3212788173 hasConcept C154945302 @default.
- W3212788173 hasConcept C159985019 @default.
- W3212788173 hasConcept C162307627 @default.
- W3212788173 hasConcept C162324750 @default.
- W3212788173 hasConcept C177148314 @default.
- W3212788173 hasConcept C180016635 @default.
- W3212788173 hasConcept C187736073 @default.
- W3212788173 hasConcept C192562407 @default.
- W3212788173 hasConcept C206345919 @default.
- W3212788173 hasConcept C2780451532 @default.
- W3212788173 hasConcept C2781002164 @default.
- W3212788173 hasConcept C31258907 @default.
- W3212788173 hasConcept C33923547 @default.
- W3212788173 hasConcept C41008148 @default.
- W3212788173 hasConcept C44291984 @default.
- W3212788173 hasConcept C97541855 @default.
- W3212788173 hasConcept C98045186 @default.
- W3212788173 hasConceptScore W3212788173C111919701 @default.
- W3212788173 hasConceptScore W3212788173C119857082 @default.
- W3212788173 hasConceptScore W3212788173C134306372 @default.
- W3212788173 hasConceptScore W3212788173C137293760 @default.
- W3212788173 hasConceptScore W3212788173C154945302 @default.
- W3212788173 hasConceptScore W3212788173C159985019 @default.
- W3212788173 hasConceptScore W3212788173C162307627 @default.
- W3212788173 hasConceptScore W3212788173C162324750 @default.
- W3212788173 hasConceptScore W3212788173C177148314 @default.
- W3212788173 hasConceptScore W3212788173C180016635 @default.
- W3212788173 hasConceptScore W3212788173C187736073 @default.
- W3212788173 hasConceptScore W3212788173C192562407 @default.
- W3212788173 hasConceptScore W3212788173C206345919 @default.
- W3212788173 hasConceptScore W3212788173C2780451532 @default.
- W3212788173 hasConceptScore W3212788173C2781002164 @default.
- W3212788173 hasConceptScore W3212788173C31258907 @default.
- W3212788173 hasConceptScore W3212788173C33923547 @default.
- W3212788173 hasConceptScore W3212788173C41008148 @default.
- W3212788173 hasConceptScore W3212788173C44291984 @default.
- W3212788173 hasConceptScore W3212788173C97541855 @default.
- W3212788173 hasConceptScore W3212788173C98045186 @default.
- W3212788173 hasFunder F4320320696 @default.
- W3212788173 hasFunder F4320322922 @default.
- W3212788173 hasLocation W32127881731 @default.
- W3212788173 hasOpenAccess W3212788173 @default.
- W3212788173 hasPrimaryLocation W32127881731 @default.
- W3212788173 hasRelatedWork W2979869797 @default.
- W3212788173 hasRelatedWork W3035417664 @default.
- W3212788173 hasRelatedWork W3090436287 @default.
- W3212788173 hasRelatedWork W3199608561 @default.
- W3212788173 hasRelatedWork W4226029999 @default.
- W3212788173 hasRelatedWork W4285338583 @default.
- W3212788173 hasRelatedWork W4287647350 @default.
- W3212788173 hasRelatedWork W4308664185 @default.
- W3212788173 hasRelatedWork W4319083788 @default.
- W3212788173 hasRelatedWork W4319309271 @default.
- W3212788173 hasVolume "487" @default.
- W3212788173 isParatext "false" @default.
- W3212788173 isRetracted "false" @default.
- W3212788173 magId "3212788173" @default.
- W3212788173 workType "article" @default.