Matches in SemOpenAlex for { <https://semopenalex.org/work/W4221086658> ?p ?o ?g. }
- W4221086658 abstract "This article examines the basis of Natural Language Understanding of transformer based language models, such as BERT. It does this through a case study on idiom token classification. We use idiom token identification as a basis for our analysis because of the variety of information types that have previously been explored in the literature for this task, including: topic, lexical, and syntactic features. This variety of relevant information types means that the task of idiom token identification enables us to explore the forms of linguistic information that a BERT language model captures and encodes in its representations. The core of this article presents three experiments. The first experiment analyzes the effectiveness of BERT sentence embeddings for creating a general idiom token identification model and the results indicate that the BERT sentence embeddings outperform Skip-Thought. In the second and third experiment we use the game theory concept of Shapley Values to rank the usefulness of individual idiomatic expressions for model training and use this ranking to analyse the type of information that the model finds useful. We find that a combination of idiom-intrinsic and topic-based properties contribute to an expression's usefulness in idiom token identification. Overall our results indicate that BERT efficiently encodes a variety of information from topic, through lexical and syntactic information. Based on these results we argue that notwithstanding recent criticisms of language model based semantics, the ability of BERT to efficiently encode a variety of linguistic information types does represent a significant step forward in natural language understanding." @default.
- W4221086658 created "2022-04-03" @default.
- W4221086658 creator A5021204772 @default.
- W4221086658 creator A5057328421 @default.
- W4221086658 creator A5058781219 @default.
- W4221086658 date "2022-03-14" @default.
- W4221086658 modified "2023-10-17" @default.
- W4221086658 title "Shapley Idioms: Analysing BERT Sentence Embeddings for General Idiom Token Identification" @default.
- W4221086658 cites W1498763386 @default.
- W4221086658 cites W172541218 @default.
- W4221086658 cites W1972686387 @default.
- W4221086658 cites W1997721713 @default.
- W4221086658 cites W2045705339 @default.
- W4221086658 cites W2123215530 @default.
- W4221086658 cites W2141179513 @default.
- W4221086658 cites W2147152072 @default.
- W4221086658 cites W2517996181 @default.
- W4221086658 cites W2770760637 @default.
- W4221086658 cites W2942649361 @default.
- W4221086658 cites W2946417913 @default.
- W4221086658 cites W2964204621 @default.
- W4221086658 cites W2964274713 @default.
- W4221086658 cites W2971016963 @default.
- W4221086658 cites W2997591727 @default.
- W4221086658 cites W3004346089 @default.
- W4221086658 cites W3034723486 @default.
- W4221086658 cites W3090831419 @default.
- W4221086658 cites W3118485687 @default.
- W4221086658 cites W3183582945 @default.
- W4221086658 cites W4245888765 @default.
- W4221086658 doi "https://doi.org/10.3389/frai.2022.813967" @default.
- W4221086658 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/35360661" @default.
- W4221086658 hasPublicationYear "2022" @default.
- W4221086658 type Work @default.
- W4221086658 citedByCount "2" @default.
- W4221086658 countsByYear W42210866582022 @default.
- W4221086658 countsByYear W42210866582023 @default.
- W4221086658 crossrefType "journal-article" @default.
- W4221086658 hasAuthorship W4221086658A5021204772 @default.
- W4221086658 hasAuthorship W4221086658A5057328421 @default.
- W4221086658 hasAuthorship W4221086658A5058781219 @default.
- W4221086658 hasBestOaLocation W42210866581 @default.
- W4221086658 hasConcept C104317684 @default.
- W4221086658 hasConcept C114614502 @default.
- W4221086658 hasConcept C116834253 @default.
- W4221086658 hasConcept C136197465 @default.
- W4221086658 hasConcept C137293760 @default.
- W4221086658 hasConcept C138885662 @default.
- W4221086658 hasConcept C154945302 @default.
- W4221086658 hasConcept C164226766 @default.
- W4221086658 hasConcept C185592680 @default.
- W4221086658 hasConcept C195324797 @default.
- W4221086658 hasConcept C204321447 @default.
- W4221086658 hasConcept C2777530160 @default.
- W4221086658 hasConcept C2779439875 @default.
- W4221086658 hasConcept C33923547 @default.
- W4221086658 hasConcept C38652104 @default.
- W4221086658 hasConcept C41008148 @default.
- W4221086658 hasConcept C41895202 @default.
- W4221086658 hasConcept C44291984 @default.
- W4221086658 hasConcept C48145219 @default.
- W4221086658 hasConcept C55493867 @default.
- W4221086658 hasConcept C59822182 @default.
- W4221086658 hasConcept C66746571 @default.
- W4221086658 hasConcept C86803240 @default.
- W4221086658 hasConceptScore W4221086658C104317684 @default.
- W4221086658 hasConceptScore W4221086658C114614502 @default.
- W4221086658 hasConceptScore W4221086658C116834253 @default.
- W4221086658 hasConceptScore W4221086658C136197465 @default.
- W4221086658 hasConceptScore W4221086658C137293760 @default.
- W4221086658 hasConceptScore W4221086658C138885662 @default.
- W4221086658 hasConceptScore W4221086658C154945302 @default.
- W4221086658 hasConceptScore W4221086658C164226766 @default.
- W4221086658 hasConceptScore W4221086658C185592680 @default.
- W4221086658 hasConceptScore W4221086658C195324797 @default.
- W4221086658 hasConceptScore W4221086658C204321447 @default.
- W4221086658 hasConceptScore W4221086658C2777530160 @default.
- W4221086658 hasConceptScore W4221086658C2779439875 @default.
- W4221086658 hasConceptScore W4221086658C33923547 @default.
- W4221086658 hasConceptScore W4221086658C38652104 @default.
- W4221086658 hasConceptScore W4221086658C41008148 @default.
- W4221086658 hasConceptScore W4221086658C41895202 @default.
- W4221086658 hasConceptScore W4221086658C44291984 @default.
- W4221086658 hasConceptScore W4221086658C48145219 @default.
- W4221086658 hasConceptScore W4221086658C55493867 @default.
- W4221086658 hasConceptScore W4221086658C59822182 @default.
- W4221086658 hasConceptScore W4221086658C66746571 @default.
- W4221086658 hasConceptScore W4221086658C86803240 @default.
- W4221086658 hasFunder F4320325671 @default.
- W4221086658 hasLocation W42210866581 @default.
- W4221086658 hasLocation W42210866582 @default.
- W4221086658 hasLocation W42210866583 @default.
- W4221086658 hasLocation W42210866584 @default.
- W4221086658 hasLocation W42210866585 @default.
- W4221086658 hasOpenAccess W4221086658 @default.
- W4221086658 hasPrimaryLocation W42210866581 @default.
- W4221086658 hasRelatedWork W159132833 @default.
- W4221086658 hasRelatedWork W2033261979 @default.
- W4221086658 hasRelatedWork W20999564 @default.
- W4221086658 hasRelatedWork W2411652523 @default.