Matches in SemOpenAlex for { <https://semopenalex.org/work/W2950382646> ?p ?o ?g. }
- W2950382646 abstract "While paragraph embedding models are remarkably effective for downstream classification tasks, what they learn and encode into a single vector remains opaque. In this paper, we investigate a state-of-the-art paragraph embedding method proposed by Zhang et al. (2017) and discover that it cannot reliably tell whether a given sentence occurs in the input paragraph or not. We formulate a sentence content task to probe for this basic linguistic property and find that even a much simpler bag-of-words method has no trouble solving it. This result motivates us to replace the reconstruction-based objective of Zhang et al. (2017) with our sentence content probe objective in a semi-supervised setting. Despite its simplicity, our objective improves over paragraph reconstruction in terms of (1) downstream classification accuracies on benchmark datasets, (2) faster training, and (3) better generalization ability." @default.
- W2950382646 created "2019-06-27" @default.
- W2950382646 creator A5035216205 @default.
- W2950382646 creator A5082767919 @default.
- W2950382646 date "2019-01-01" @default.
- W2950382646 modified "2023-10-16" @default.
- W2950382646 title "Encouraging Paragraph Embeddings to Remember Sentence Identity Improves Classification" @default.
- W2950382646 cites W1486649854 @default.
- W2950382646 cites W2064675550 @default.
- W2950382646 cites W2095705004 @default.
- W2950382646 cites W2113459411 @default.
- W2950382646 cites W2115613106 @default.
- W2950382646 cites W2130942839 @default.
- W2950382646 cites W2131744502 @default.
- W2950382646 cites W2153579005 @default.
- W2950382646 cites W2250473257 @default.
- W2950382646 cites W2250539671 @default.
- W2950382646 cites W2563574619 @default.
- W2950382646 cites W2605717780 @default.
- W2950382646 cites W2610858497 @default.
- W2950382646 cites W2761988601 @default.
- W2950382646 cites W2794557536 @default.
- W2950382646 cites W2962739339 @default.
- W2950382646 cites W2962776659 @default.
- W2950382646 cites W2962777840 @default.
- W2950382646 cites W2962800561 @default.
- W2950382646 cites W2963012544 @default.
- W2950382646 cites W2963026768 @default.
- W2950382646 cites W2963090765 @default.
- W2950382646 cites W2963341956 @default.
- W2950382646 cites W2963756346 @default.
- W2950382646 cites W2963899155 @default.
- W2950382646 cites W2963918774 @default.
- W2950382646 cites W2964028602 @default.
- W2950382646 cites W2964121744 @default.
- W2950382646 cites W2964204621 @default.
- W2950382646 cites W941230081 @default.
- W2950382646 doi "https://doi.org/10.18653/v1/p19-1638" @default.
- W2950382646 hasPublicationYear "2019" @default.
- W2950382646 type Work @default.
- W2950382646 sameAs 2950382646 @default.
- W2950382646 citedByCount "3" @default.
- W2950382646 countsByYear W29503826462020 @default.
- W2950382646 countsByYear W29503826462023 @default.
- W2950382646 crossrefType "proceedings-article" @default.
- W2950382646 hasAuthorship W2950382646A5035216205 @default.
- W2950382646 hasAuthorship W2950382646A5082767919 @default.
- W2950382646 hasBestOaLocation W29503826461 @default.
- W2950382646 hasConcept C111472728 @default.
- W2950382646 hasConcept C13280743 @default.
- W2950382646 hasConcept C134306372 @default.
- W2950382646 hasConcept C136764020 @default.
- W2950382646 hasConcept C138885662 @default.
- W2950382646 hasConcept C153180895 @default.
- W2950382646 hasConcept C154945302 @default.
- W2950382646 hasConcept C162324750 @default.
- W2950382646 hasConcept C177148314 @default.
- W2950382646 hasConcept C185798385 @default.
- W2950382646 hasConcept C187736073 @default.
- W2950382646 hasConcept C189950617 @default.
- W2950382646 hasConcept C204321447 @default.
- W2950382646 hasConcept C205649164 @default.
- W2950382646 hasConcept C2776372474 @default.
- W2950382646 hasConcept C2777206241 @default.
- W2950382646 hasConcept C2777530160 @default.
- W2950382646 hasConcept C2780451532 @default.
- W2950382646 hasConcept C28490314 @default.
- W2950382646 hasConcept C33923547 @default.
- W2950382646 hasConcept C41008148 @default.
- W2950382646 hasConcept C41608201 @default.
- W2950382646 hasConceptScore W2950382646C111472728 @default.
- W2950382646 hasConceptScore W2950382646C13280743 @default.
- W2950382646 hasConceptScore W2950382646C134306372 @default.
- W2950382646 hasConceptScore W2950382646C136764020 @default.
- W2950382646 hasConceptScore W2950382646C138885662 @default.
- W2950382646 hasConceptScore W2950382646C153180895 @default.
- W2950382646 hasConceptScore W2950382646C154945302 @default.
- W2950382646 hasConceptScore W2950382646C162324750 @default.
- W2950382646 hasConceptScore W2950382646C177148314 @default.
- W2950382646 hasConceptScore W2950382646C185798385 @default.
- W2950382646 hasConceptScore W2950382646C187736073 @default.
- W2950382646 hasConceptScore W2950382646C189950617 @default.
- W2950382646 hasConceptScore W2950382646C204321447 @default.
- W2950382646 hasConceptScore W2950382646C205649164 @default.
- W2950382646 hasConceptScore W2950382646C2776372474 @default.
- W2950382646 hasConceptScore W2950382646C2777206241 @default.
- W2950382646 hasConceptScore W2950382646C2777530160 @default.
- W2950382646 hasConceptScore W2950382646C2780451532 @default.
- W2950382646 hasConceptScore W2950382646C28490314 @default.
- W2950382646 hasConceptScore W2950382646C33923547 @default.
- W2950382646 hasConceptScore W2950382646C41008148 @default.
- W2950382646 hasConceptScore W2950382646C41608201 @default.
- W2950382646 hasLocation W29503826461 @default.
- W2950382646 hasLocation W29503826462 @default.
- W2950382646 hasOpenAccess W2950382646 @default.
- W2950382646 hasPrimaryLocation W29503826461 @default.
- W2950382646 hasRelatedWork W1601017932 @default.
- W2950382646 hasRelatedWork W2013781749 @default.
- W2950382646 hasRelatedWork W2070147364 @default.
- W2950382646 hasRelatedWork W2081647779 @default.