Matches in SemOpenAlex for { <https://semopenalex.org/work/W2972969579> ?p ?o ?g. }
- W2972969579 abstract "Contextual word representations, typically trained on unstructured, unlabeled text, do not contain any explicit grounding to real world entities and are often unable to remember facts about those entities. We propose a general method to embed multiple knowledge bases (KBs) into large scale models, and thereby enhance their representations with structured, human-curated knowledge. For each KB, we first use an integrated entity linker to retrieve relevant entity embeddings, then update contextual word representations via a form of word-to-entity attention. In contrast to previous approaches, the entity linkers and self-supervised language modeling objective are jointly trained end-to-end in a multitask setting that combines a small amount of entity linking supervision with a large amount of raw text. After integrating WordNet and a subset of Wikipedia into BERT, the knowledge enhanced BERT (KnowBert) demonstrates improved perplexity, ability to recall facts as measured in a probing task and downstream performance on relationship extraction, entity typing, and word sense disambiguation. KnowBert's runtime is comparable to BERT's and it scales to large KBs." @default.
- W2972969579 created "2019-09-19" @default.
- W2972969579 creator A5005779128 @default.
- W2972969579 creator A5007903277 @default.
- W2972969579 creator A5016370551 @default.
- W2972969579 creator A5017963313 @default.
- W2972969579 creator A5031742397 @default.
- W2972969579 creator A5049944832 @default.
- W2972969579 creator A5088517824 @default.
- W2972969579 date "2019-09-09" @default.
- W2972969579 modified "2023-09-23" @default.
- W2972969579 title "Knowledge Enhanced Contextual Word Representations" @default.
- W2972969579 cites W11298561 @default.
- W2972969579 cites W1566289585 @default.
- W2972969579 cites W1713614699 @default.
- W2972969579 cites W2026810221 @default.
- W2972969579 cites W2035717317 @default.
- W2972969579 cites W205829674 @default.
- W2972969579 cites W2080133951 @default.
- W2972969579 cites W2081580037 @default.
- W2972969579 cites W2099779943 @default.
- W2972969579 cites W2117130368 @default.
- W2972969579 cites W2123142779 @default.
- W2972969579 cites W2153579005 @default.
- W2972969579 cites W2158028897 @default.
- W2972969579 cites W2159583324 @default.
- W2972969579 cites W2184957013 @default.
- W2972969579 cites W2250539671 @default.
- W2972969579 cites W2250750514 @default.
- W2972969579 cites W2251537235 @default.
- W2972969579 cites W2283196293 @default.
- W2972969579 cites W2476140796 @default.
- W2972969579 cites W2499696929 @default.
- W2972969579 cites W2507974895 @default.
- W2972969579 cites W2513378248 @default.
- W2972969579 cites W2525778437 @default.
- W2972969579 cites W2612773933 @default.
- W2972969579 cites W2740782137 @default.
- W2972969579 cites W2741075451 @default.
- W2972969579 cites W2757205734 @default.
- W2972969579 cites W2757369719 @default.
- W2972969579 cites W2759211898 @default.
- W2972969579 cites W2798416089 @default.
- W2972969579 cites W2888236192 @default.
- W2972969579 cites W2890961898 @default.
- W2972969579 cites W2892094955 @default.
- W2972969579 cites W2892280852 @default.
- W2972969579 cites W2936215830 @default.
- W2972969579 cites W2950577311 @default.
- W2972969579 cites W2951048068 @default.
- W2972969579 cites W2951077644 @default.
- W2972969579 cites W2951561177 @default.
- W2972969579 cites W2953044594 @default.
- W2972969579 cites W2953356739 @default.
- W2972969579 cites W2956105246 @default.
- W2972969579 cites W2962739339 @default.
- W2972969579 cites W2962784628 @default.
- W2972969579 cites W2962891712 @default.
- W2972969579 cites W2963026768 @default.
- W2972969579 cites W2963090765 @default.
- W2972969579 cites W2963290255 @default.
- W2972969579 cites W2963341956 @default.
- W2972969579 cites W2963403868 @default.
- W2972969579 cites W2963432357 @default.
- W2972969579 cites W2963534707 @default.
- W2972969579 cites W2963613359 @default.
- W2972969579 cites W2963756346 @default.
- W2972969579 cites W2963829073 @default.
- W2972969579 cites W2963907426 @default.
- W2972969579 cites W2963925965 @default.
- W2972969579 cites W2964022985 @default.
- W2972969579 cites W2964116313 @default.
- W2972969579 cites W2964121744 @default.
- W2972969579 cites W2964207259 @default.
- W2972969579 cites W2964222246 @default.
- W2972969579 cites W2990704537 @default.
- W2972969579 cites W3103296573 @default.
- W2972969579 cites W3106003309 @default.
- W2972969579 hasPublicationYear "2019" @default.
- W2972969579 type Work @default.
- W2972969579 sameAs 2972969579 @default.
- W2972969579 citedByCount "12" @default.
- W2972969579 countsByYear W29729695792020 @default.
- W2972969579 countsByYear W29729695792021 @default.
- W2972969579 countsByYear W29729695792022 @default.
- W2972969579 crossrefType "posted-content" @default.
- W2972969579 hasAuthorship W2972969579A5005779128 @default.
- W2972969579 hasAuthorship W2972969579A5007903277 @default.
- W2972969579 hasAuthorship W2972969579A5016370551 @default.
- W2972969579 hasAuthorship W2972969579A5017963313 @default.
- W2972969579 hasAuthorship W2972969579A5031742397 @default.
- W2972969579 hasAuthorship W2972969579A5049944832 @default.
- W2972969579 hasAuthorship W2972969579A5088517824 @default.
- W2972969579 hasConcept C100279451 @default.
- W2972969579 hasConcept C137293760 @default.
- W2972969579 hasConcept C138885662 @default.
- W2972969579 hasConcept C154945302 @default.
- W2972969579 hasConcept C157659113 @default.
- W2972969579 hasConcept C162324750 @default.
- W2972969579 hasConcept C187736073 @default.