Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287671657> ?p ?o ?g. }
Showing items 1 to 53 of
53
with 100 items per page.
- W4287671657 abstract "In this paper we introduce a method for visually analyzing contextualized embeddings produced by deep neural network-based language models. Our approach is inspired by linguistic probes for natural language processing, where tasks are designed to probe language models for linguistic structure, such as parts-of-speech and named entities. These approaches are largely confirmatory, however, only enabling a user to test for information known a priori. In this work, we eschew supervised probing tasks, and advocate for unsupervised probes, coupled with visual exploration techniques, to assess what is learned by language models. Specifically, we cluster contextualized embeddings produced from a large text corpus, and introduce a visualization design based on this clustering and textual structure - cluster co-occurrences, cluster spans, and cluster-word membership - to help elicit the functionality of, and relationship between, individual clusters. User feedback highlights the benefits of our design in discovering different types of linguistic structures." @default.
- W4287671657 created "2022-07-25" @default.
- W4287671657 creator A5060051655 @default.
- W4287671657 date "2020-09-05" @default.
- W4287671657 modified "2023-09-24" @default.
- W4287671657 title "Visually Analyzing Contextualized Embeddings" @default.
- W4287671657 doi "https://doi.org/10.48550/arxiv.2009.02554" @default.
- W4287671657 hasPublicationYear "2020" @default.
- W4287671657 type Work @default.
- W4287671657 citedByCount "0" @default.
- W4287671657 crossrefType "posted-content" @default.
- W4287671657 hasAuthorship W4287671657A5060051655 @default.
- W4287671657 hasBestOaLocation W42876716571 @default.
- W4287671657 hasConcept C111472728 @default.
- W4287671657 hasConcept C137293760 @default.
- W4287671657 hasConcept C138885662 @default.
- W4287671657 hasConcept C154945302 @default.
- W4287671657 hasConcept C195324797 @default.
- W4287671657 hasConcept C204321447 @default.
- W4287671657 hasConcept C36464697 @default.
- W4287671657 hasConcept C41008148 @default.
- W4287671657 hasConcept C41895202 @default.
- W4287671657 hasConcept C73555534 @default.
- W4287671657 hasConcept C75553542 @default.
- W4287671657 hasConcept C90805587 @default.
- W4287671657 hasConceptScore W4287671657C111472728 @default.
- W4287671657 hasConceptScore W4287671657C137293760 @default.
- W4287671657 hasConceptScore W4287671657C138885662 @default.
- W4287671657 hasConceptScore W4287671657C154945302 @default.
- W4287671657 hasConceptScore W4287671657C195324797 @default.
- W4287671657 hasConceptScore W4287671657C204321447 @default.
- W4287671657 hasConceptScore W4287671657C36464697 @default.
- W4287671657 hasConceptScore W4287671657C41008148 @default.
- W4287671657 hasConceptScore W4287671657C41895202 @default.
- W4287671657 hasConceptScore W4287671657C73555534 @default.
- W4287671657 hasConceptScore W4287671657C75553542 @default.
- W4287671657 hasConceptScore W4287671657C90805587 @default.
- W4287671657 hasLocation W42876716571 @default.
- W4287671657 hasOpenAccess W4287671657 @default.
- W4287671657 hasPrimaryLocation W42876716571 @default.
- W4287671657 hasRelatedWork W10423536 @default.
- W4287671657 hasRelatedWork W12256131 @default.
- W4287671657 hasRelatedWork W13780460 @default.
- W4287671657 hasRelatedWork W14230040 @default.
- W4287671657 hasRelatedWork W17684 @default.
- W4287671657 hasRelatedWork W4615079 @default.
- W4287671657 hasRelatedWork W4629839 @default.
- W4287671657 hasRelatedWork W5688477 @default.
- W4287671657 hasRelatedWork W7636100 @default.
- W4287671657 hasRelatedWork W8053366 @default.
- W4287671657 isParatext "false" @default.
- W4287671657 isRetracted "false" @default.
- W4287671657 workType "article" @default.