FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3136346046> ?p ?o ?g. }

• W3136346046 endingPage "102563" @default.
• W3136346046 startingPage "102563" @default.
• W3136346046 abstract "This paper presents DWIE, the ‘Deutsche Welle corpus for Information Extraction’, a newly created multi-task dataset that combines four main Information Extraction (IE) annotation subtasks: (i) Named Entity Recognition (NER), (ii) Coreference Resolution, (iii) Relation Extraction (RE), and (iv) Entity Linking. DWIE is conceived as an entity-centric dataset that describes interactions and properties of conceptual entities on the level of the complete document. This contrasts with currently dominant mention-driven approaches that start from the detection and classification of named entity mentions in individual sentences. Further, DWIE presented two main challenges when building and evaluating IE models for it. First, the use of traditional mention-level evaluation metrics for NER and RE tasks on entity-centric DWIE dataset can result in measurements dominated by predictions on more frequently mentioned entities. We tackle this issue by proposing a new entity-driven metric that takes into account the number of mentions that compose each of the predicted and ground truth entities. Second, the document-level multi-task annotations require the models to transfer information between entity mentions located in different parts of the document, as well as between different tasks, in a joint learning setting. To realize this, we propose to use graph-based neural message passing techniques between document-level mention spans. Our experiments show an improvement of up to 5.5 F1 percentage points when incorporating neural graph propagation into our joint model. This demonstrates DWIE’s potential to stimulate further research in graph neural networks for representation learning in multi-task IE. We make DWIE publicly available at https://github.com/klimzaporojets/DWIE." @default.
• W3136346046 created "2021-03-29" @default.
• W3136346046 creator A5004214531 @default.
• W3136346046 creator A5011614450 @default.
• W3136346046 creator A5075509168 @default.
• W3136346046 creator A5084742757 @default.
• W3136346046 date "2021-07-01" @default.
• W3136346046 modified "2023-10-14" @default.
• W3136346046 title "DWIE: An entity-centric dataset for multi-task document-level information extraction" @default.
• W3136346046 cites W1604644367 @default.
• W3136346046 cites W2053154970 @default.
• W3136346046 cites W2088911157 @default.
• W3136346046 cites W2098345921 @default.
• W3136346046 cites W2099779943 @default.
• W3136346046 cites W2115352105 @default.
• W3136346046 cites W2116341502 @default.
• W3136346046 cites W2133943399 @default.
• W3136346046 cites W2134033474 @default.
• W3136346046 cites W2163107094 @default.
• W3136346046 cites W2164777277 @default.
• W3136346046 cites W2250539671 @default.
• W3136346046 cites W2250576901 @default.
• W3136346046 cites W2250668331 @default.
• W3136346046 cites W2251064706 @default.
• W3136346046 cites W2251628379 @default.
• W3136346046 cites W2296283641 @default.
• W3136346046 cites W2548006041 @default.
• W3136346046 cites W2619266022 @default.
• W3136346046 cites W2759056771 @default.
• W3136346046 cites W2759211898 @default.
• W3136346046 cites W2760505947 @default.
• W3136346046 cites W2788460763 @default.
• W3136346046 cites W2799125718 @default.
• W3136346046 cites W2803609931 @default.
• W3136346046 cites W2805720864 @default.
• W3136346046 cites W2891602716 @default.
• W3136346046 cites W2892094955 @default.
• W3136346046 cites W2920114910 @default.
• W3136346046 cites W2925285378 @default.
• W3136346046 cites W2935052563 @default.
• W3136346046 cites W2946558277 @default.
• W3136346046 cites W2949204161 @default.
• W3136346046 cites W2949212908 @default.
• W3136346046 cites W2949493644 @default.
• W3136346046 cites W2950970512 @default.
• W3136346046 cites W2952179106 @default.
• W3136346046 cites W2952768212 @default.
• W3136346046 cites W2962798796 @default.
• W3136346046 cites W2962902328 @default.
• W3136346046 cites W2962903510 @default.
• W3136346046 cites W2962985038 @default.
• W3136346046 cites W2963020213 @default.
• W3136346046 cites W2963087868 @default.
• W3136346046 cites W2963140597 @default.
• W3136346046 cites W2963625095 @default.
• W3136346046 cites W2963655104 @default.
• W3136346046 cites W2963670884 @default.
• W3136346046 cites W2963718112 @default.
• W3136346046 cites W2963777632 @default.
• W3136346046 cites W2963863453 @default.
• W3136346046 cites W2963869731 @default.
• W3136346046 cites W2963997908 @default.
• W3136346046 cites W2964022985 @default.
• W3136346046 cites W2964193968 @default.
• W3136346046 cites W2964222246 @default.
• W3136346046 cites W2970119519 @default.
• W3136346046 cites W2970986510 @default.
• W3136346046 cites W2984147501 @default.
• W3136346046 cites W2984452801 @default.
• W3136346046 cites W2984582583 @default.
• W3136346046 cites W2998631105 @default.
• W3136346046 cites W3035537500 @default.
• W3136346046 cites W3042179685 @default.
• W3136346046 cites W3042708284 @default.
• W3136346046 cites W3105358070 @default.
• W3136346046 cites W4294214983 @default.
• W3136346046 doi "https://doi.org/10.1016/j.ipm.2021.102563" @default.
• W3136346046 hasPublicationYear "2021" @default.
• W3136346046 type Work @default.
• W3136346046 sameAs 3136346046 @default.
• W3136346046 citedByCount "20" @default.
• W3136346046 countsByYear W31363460462021 @default.
• W3136346046 countsByYear W31363460462022 @default.
• W3136346046 countsByYear W31363460462023 @default.
• W3136346046 crossrefType "journal-article" @default.
• W3136346046 hasAuthorship W3136346046A5004214531 @default.
• W3136346046 hasAuthorship W3136346046A5011614450 @default.
• W3136346046 hasAuthorship W3136346046A5075509168 @default.
• W3136346046 hasAuthorship W3136346046A5084742757 @default.
• W3136346046 hasBestOaLocation W31363460462 @default.
• W3136346046 hasConcept C132525143 @default.
• W3136346046 hasConcept C138268822 @default.
• W3136346046 hasConcept C146849305 @default.
• W3136346046 hasConcept C153604712 @default.
• W3136346046 hasConcept C154945302 @default.
• W3136346046 hasConcept C162324750 @default.
• W3136346046 hasConcept C176217482 @default.
• W3136346046 hasConcept C187736073 @default.

• next