Matches in SemOpenAlex for { <https://semopenalex.org/work/W4226252996> ?p ?o ?g. }
Showing items 1 to 97 of
97
with 100 items per page.
- W4226252996 endingPage "233" @default.
- W4226252996 startingPage "222" @default.
- W4226252996 abstract "Recent advances on prompt-tuning cast few-shot classification tasks as a masked language modeling problem. By wrapping input into a template and using a verbalizer which constructs a mapping between label space and label word space, prompt-tuning can achieve excellent results in few-shot scenarios. However, typical prompt-tuning needs a manually designed verbalizer which requires domain expertise and human efforts. And the insufficient label space may introduce considerable bias into the results. In this paper, we focus on eliciting knowledge from pretrained language models and propose a prototypical prompt verbalizer for prompt-tuning. Labels are represented by prototypical embeddings in the feature space rather than by discrete words. The distances between the embedding at the masked position of input and prototypical embeddings are used as classification criterion. To address the problem of random initialization of parameters in zero-shot settings, we elicit knowledge from pretrained language models to form initial prototypical embeddings. Our method optimizes models by contrastive learning. Extensive experimental results on several many-class text classification datasets with low-resource settings demonstrate the effectiveness of our approach compared with other verbalizer construction methods. Our implementation is at https://github.com/Ydongd/prototypical-prompt-verbalizer ." @default.
- W4226252996 created "2022-05-05" @default.
- W4226252996 creator A5019904424 @default.
- W4226252996 creator A5059356240 @default.
- W4226252996 creator A5060667885 @default.
- W4226252996 creator A5069907865 @default.
- W4226252996 creator A5078949281 @default.
- W4226252996 date "2022-01-01" @default.
- W4226252996 modified "2023-10-13" @default.
- W4226252996 title "Eliciting Knowledge from Pretrained Language Models for Prototypical Prompt Verbalizer" @default.
- W4226252996 cites W1552847225 @default.
- W4226252996 cites W2798991696 @default.
- W4226252996 cites W3040834782 @default.
- W4226252996 cites W3096580779 @default.
- W4226252996 cites W3102659883 @default.
- W4226252996 cites W3153427360 @default.
- W4226252996 cites W3156636935 @default.
- W4226252996 cites W3174784402 @default.
- W4226252996 cites W3188542058 @default.
- W4226252996 cites W3210129272 @default.
- W4226252996 doi "https://doi.org/10.1007/978-3-031-15931-2_19" @default.
- W4226252996 hasPublicationYear "2022" @default.
- W4226252996 type Work @default.
- W4226252996 citedByCount "2" @default.
- W4226252996 countsByYear W42262529962023 @default.
- W4226252996 crossrefType "book-chapter" @default.
- W4226252996 hasAuthorship W4226252996A5019904424 @default.
- W4226252996 hasAuthorship W4226252996A5059356240 @default.
- W4226252996 hasAuthorship W4226252996A5060667885 @default.
- W4226252996 hasAuthorship W4226252996A5069907865 @default.
- W4226252996 hasAuthorship W4226252996A5078949281 @default.
- W4226252996 hasBestOaLocation W42262529962 @default.
- W4226252996 hasConcept C111919701 @default.
- W4226252996 hasConcept C114466953 @default.
- W4226252996 hasConcept C119857082 @default.
- W4226252996 hasConcept C120665830 @default.
- W4226252996 hasConcept C121332964 @default.
- W4226252996 hasConcept C134306372 @default.
- W4226252996 hasConcept C137293760 @default.
- W4226252996 hasConcept C138885662 @default.
- W4226252996 hasConcept C154945302 @default.
- W4226252996 hasConcept C192209626 @default.
- W4226252996 hasConcept C199360897 @default.
- W4226252996 hasConcept C204321447 @default.
- W4226252996 hasConcept C2776401178 @default.
- W4226252996 hasConcept C2777212361 @default.
- W4226252996 hasConcept C2777462759 @default.
- W4226252996 hasConcept C2778572836 @default.
- W4226252996 hasConcept C33923547 @default.
- W4226252996 hasConcept C36503486 @default.
- W4226252996 hasConcept C41008148 @default.
- W4226252996 hasConcept C41608201 @default.
- W4226252996 hasConcept C41895202 @default.
- W4226252996 hasConcept C83665646 @default.
- W4226252996 hasConcept C90805587 @default.
- W4226252996 hasConceptScore W4226252996C111919701 @default.
- W4226252996 hasConceptScore W4226252996C114466953 @default.
- W4226252996 hasConceptScore W4226252996C119857082 @default.
- W4226252996 hasConceptScore W4226252996C120665830 @default.
- W4226252996 hasConceptScore W4226252996C121332964 @default.
- W4226252996 hasConceptScore W4226252996C134306372 @default.
- W4226252996 hasConceptScore W4226252996C137293760 @default.
- W4226252996 hasConceptScore W4226252996C138885662 @default.
- W4226252996 hasConceptScore W4226252996C154945302 @default.
- W4226252996 hasConceptScore W4226252996C192209626 @default.
- W4226252996 hasConceptScore W4226252996C199360897 @default.
- W4226252996 hasConceptScore W4226252996C204321447 @default.
- W4226252996 hasConceptScore W4226252996C2776401178 @default.
- W4226252996 hasConceptScore W4226252996C2777212361 @default.
- W4226252996 hasConceptScore W4226252996C2777462759 @default.
- W4226252996 hasConceptScore W4226252996C2778572836 @default.
- W4226252996 hasConceptScore W4226252996C33923547 @default.
- W4226252996 hasConceptScore W4226252996C36503486 @default.
- W4226252996 hasConceptScore W4226252996C41008148 @default.
- W4226252996 hasConceptScore W4226252996C41608201 @default.
- W4226252996 hasConceptScore W4226252996C41895202 @default.
- W4226252996 hasConceptScore W4226252996C83665646 @default.
- W4226252996 hasConceptScore W4226252996C90805587 @default.
- W4226252996 hasLocation W42262529961 @default.
- W4226252996 hasLocation W42262529962 @default.
- W4226252996 hasOpenAccess W4226252996 @default.
- W4226252996 hasPrimaryLocation W42262529961 @default.
- W4226252996 hasRelatedWork W2887872604 @default.
- W4226252996 hasRelatedWork W2896411932 @default.
- W4226252996 hasRelatedWork W2944092870 @default.
- W4226252996 hasRelatedWork W2949267551 @default.
- W4226252996 hasRelatedWork W3031118452 @default.
- W4226252996 hasRelatedWork W3102086967 @default.
- W4226252996 hasRelatedWork W3184365942 @default.
- W4226252996 hasRelatedWork W33598619 @default.
- W4226252996 hasRelatedWork W4312304159 @default.
- W4226252996 hasRelatedWork W4378469273 @default.
- W4226252996 isParatext "false" @default.
- W4226252996 isRetracted "false" @default.
- W4226252996 workType "book-chapter" @default.