Matches in SemOpenAlex for { <https://semopenalex.org/work/W3175621180> ?p ?o ?g. }
Showing items 1 to 93 of
93
with 100 items per page.
- W3175621180 abstract "The major paradigm of applying a pre-trained language model to downstream tasks is to fine-tune it on labeled task data, which often suffers instability and low performance when the labeled examples are scarce.~One way to alleviate this problem is to apply post-training on unlabeled task data before fine-tuning, adapting the pre-trained model to target domains by contrastive learning that considers either token-level or sequence-level similarity. Inspired by the success of sequence masking, we argue that both token-level and sequence-level similarities can be captured with a pair of masked sequences.~Therefore, we propose complementary random masking (CRM) to generate a pair of masked sequences from an input sequence for sequence-level contrastive learning and then develop contrastive masked language modeling (CMLM) for post-training to integrate both token-level and sequence-level contrastive learnings.~Empirical results show that CMLM surpasses several recent post-training methods in few-shot settings without the need for data augmentation." @default.
- W3175621180 created "2021-07-05" @default.
- W3175621180 creator A5040062188 @default.
- W3175621180 creator A5041588125 @default.
- W3175621180 creator A5075546777 @default.
- W3175621180 date "2021-01-01" @default.
- W3175621180 modified "2023-10-05" @default.
- W3175621180 title "Bi-Granularity Contrastive Learning for Post-Training in Few-Shot Scene" @default.
- W3175621180 cites W2108598243 @default.
- W3175621180 cites W2138621090 @default.
- W3175621180 cites W2889326796 @default.
- W3175621180 cites W2898700502 @default.
- W3175621180 cites W2908510526 @default.
- W3175621180 cites W2925618549 @default.
- W3175621180 cites W2963310665 @default.
- W3175621180 cites W2963341956 @default.
- W3175621180 cites W2963403868 @default.
- W3175621180 cites W2963854351 @default.
- W3175621180 cites W2965373594 @default.
- W3175621180 cites W2995998574 @default.
- W3175621180 cites W3005700362 @default.
- W3175621180 cites W3026732421 @default.
- W3175621180 cites W3033406728 @default.
- W3175621180 cites W3034238904 @default.
- W3175621180 cites W3034978746 @default.
- W3175621180 cites W3035027743 @default.
- W3175621180 cites W3098824823 @default.
- W3175621180 cites W3100345210 @default.
- W3175621180 cites W3107668149 @default.
- W3175621180 cites W3115295967 @default.
- W3175621180 cites W3120490999 @default.
- W3175621180 cites W3122924117 @default.
- W3175621180 cites W3126074026 @default.
- W3175621180 cites W3160851104 @default.
- W3175621180 doi "https://doi.org/10.18653/v1/2021.findings-acl.151" @default.
- W3175621180 hasPublicationYear "2021" @default.
- W3175621180 type Work @default.
- W3175621180 sameAs 3175621180 @default.
- W3175621180 citedByCount "1" @default.
- W3175621180 countsByYear W31756211802023 @default.
- W3175621180 crossrefType "proceedings-article" @default.
- W3175621180 hasAuthorship W3175621180A5040062188 @default.
- W3175621180 hasAuthorship W3175621180A5041588125 @default.
- W3175621180 hasAuthorship W3175621180A5075546777 @default.
- W3175621180 hasBestOaLocation W31756211801 @default.
- W3175621180 hasConcept C127413603 @default.
- W3175621180 hasConcept C153294291 @default.
- W3175621180 hasConcept C154945302 @default.
- W3175621180 hasConcept C177774035 @default.
- W3175621180 hasConcept C191897082 @default.
- W3175621180 hasConcept C192562407 @default.
- W3175621180 hasConcept C199360897 @default.
- W3175621180 hasConcept C204321447 @default.
- W3175621180 hasConcept C205649164 @default.
- W3175621180 hasConcept C2777211547 @default.
- W3175621180 hasConcept C2778344882 @default.
- W3175621180 hasConcept C2992734406 @default.
- W3175621180 hasConcept C31972630 @default.
- W3175621180 hasConcept C41008148 @default.
- W3175621180 hasConcept C78519656 @default.
- W3175621180 hasConceptScore W3175621180C127413603 @default.
- W3175621180 hasConceptScore W3175621180C153294291 @default.
- W3175621180 hasConceptScore W3175621180C154945302 @default.
- W3175621180 hasConceptScore W3175621180C177774035 @default.
- W3175621180 hasConceptScore W3175621180C191897082 @default.
- W3175621180 hasConceptScore W3175621180C192562407 @default.
- W3175621180 hasConceptScore W3175621180C199360897 @default.
- W3175621180 hasConceptScore W3175621180C204321447 @default.
- W3175621180 hasConceptScore W3175621180C205649164 @default.
- W3175621180 hasConceptScore W3175621180C2777211547 @default.
- W3175621180 hasConceptScore W3175621180C2778344882 @default.
- W3175621180 hasConceptScore W3175621180C2992734406 @default.
- W3175621180 hasConceptScore W3175621180C31972630 @default.
- W3175621180 hasConceptScore W3175621180C41008148 @default.
- W3175621180 hasConceptScore W3175621180C78519656 @default.
- W3175621180 hasLocation W31756211801 @default.
- W3175621180 hasLocation W31756211802 @default.
- W3175621180 hasOpenAccess W3175621180 @default.
- W3175621180 hasPrimaryLocation W31756211801 @default.
- W3175621180 hasRelatedWork W1594844924 @default.
- W3175621180 hasRelatedWork W1891287906 @default.
- W3175621180 hasRelatedWork W2036807459 @default.
- W3175621180 hasRelatedWork W2354455445 @default.
- W3175621180 hasRelatedWork W2363701422 @default.
- W3175621180 hasRelatedWork W2383320921 @default.
- W3175621180 hasRelatedWork W2909382770 @default.
- W3175621180 hasRelatedWork W2981091784 @default.
- W3175621180 hasRelatedWork W3175621180 @default.
- W3175621180 hasRelatedWork W4306975314 @default.
- W3175621180 isParatext "false" @default.
- W3175621180 isRetracted "false" @default.
- W3175621180 magId "3175621180" @default.
- W3175621180 workType "article" @default.