FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2799166683> ?p ?o ?g. }

• W2799166683 abstract "Many efforts have been made to facilitate natural language processing tasks with pre-trained language models (LMs), and brought significant improvements to various applications. To fully leverage the nearly unlimited corpora and capture linguistic information of multifarious levels, large-size LMs are required; but for a specific task, only parts of these information are useful. Such large-sized LMs, even in the inference stage, may cause heavy computation workloads, making them too time-consuming for large-scale applications. Here we propose to compress bulky LMs while preserving useful information with regard to a specific task. As different layers of the model keep different information, we develop a layer selection method for model pruning using sparsity-inducing regularization. By introducing the dense connectivity, we can detach any layer without affecting others, and stretch shallow and wide LMs to be deep and narrow. In model training, LMs are learned with layer-wise dropouts for better robustness. Experiments on two benchmark datasets demonstrate the effectiveness of our method." @default.
• W2799166683 created "2018-05-07" @default.
• W2799166683 creator A5009408707 @default.
• W2799166683 creator A5019539533 @default.
• W2799166683 creator A5028523975 @default.
• W2799166683 creator A5039500313 @default.
• W2799166683 creator A5061177999 @default.
• W2799166683 date "2018-04-20" @default.
• W2799166683 modified "2023-09-27" @default.
• W2799166683 title "Efficient Contextualized Representation: Language Model Pruning for Sequence Labeling" @default.
• W2799166683 cites W1522301498 @default.
• W2799166683 cites W1690739335 @default.
• W2799166683 cites W1821462560 @default.
• W2799166683 cites W2004763266 @default.
• W2799166683 cites W2098921539 @default.
• W2799166683 cites W2160236242 @default.
• W2799166683 cites W2259472270 @default.
• W2799166683 cites W2296283641 @default.
• W2799166683 cites W2302255633 @default.
• W2799166683 cites W2319920447 @default.
• W2799166683 cites W2513419314 @default.
• W2799166683 cites W2524428287 @default.
• W2799166683 cites W2574163994 @default.
• W2799166683 cites W2606347107 @default.
• W2799166683 cites W2609130030 @default.
• W2799166683 cites W2610748790 @default.
• W2799166683 cites W2739255396 @default.
• W2799166683 cites W2756381707 @default.
• W2799166683 cites W2787560479 @default.
• W2799166683 cites W2950726992 @default.
• W2799166683 cites W2951714314 @default.
• W2799166683 cites W2952339051 @default.
• W2799166683 cites W2952729433 @default.
• W2799166683 cites W2962902328 @default.
• W2799166683 cites W2962964385 @default.
• W2799166683 cites W2963446712 @default.
• W2799166683 cites W2963625095 @default.
• W2799166683 cites W2963674932 @default.
• W2799166683 cites W2963703075 @default.
• W2799166683 cites W2963983719 @default.
• W2799166683 cites W3099701504 @default.
• W2799166683 doi "https://doi.org/10.48550/arxiv.1804.07827" @default.
• W2799166683 hasPublicationYear "2018" @default.
• W2799166683 type Work @default.
• W2799166683 sameAs 2799166683 @default.
• W2799166683 citedByCount "2" @default.
• W2799166683 countsByYear W27991666832019 @default.
• W2799166683 countsByYear W27991666832021 @default.
• W2799166683 crossrefType "posted-content" @default.
• W2799166683 hasAuthorship W2799166683A5009408707 @default.
• W2799166683 hasAuthorship W2799166683A5019539533 @default.
• W2799166683 hasAuthorship W2799166683A5028523975 @default.
• W2799166683 hasAuthorship W2799166683A5039500313 @default.
• W2799166683 hasAuthorship W2799166683A5061177999 @default.
• W2799166683 hasBestOaLocation W27991666831 @default.
• W2799166683 hasConcept C104317684 @default.
• W2799166683 hasConcept C108010975 @default.
• W2799166683 hasConcept C11413529 @default.
• W2799166683 hasConcept C119857082 @default.
• W2799166683 hasConcept C137293760 @default.
• W2799166683 hasConcept C153083717 @default.
• W2799166683 hasConcept C154945302 @default.
• W2799166683 hasConcept C162324750 @default.
• W2799166683 hasConcept C185592680 @default.
• W2799166683 hasConcept C187736073 @default.
• W2799166683 hasConcept C2776135515 @default.
• W2799166683 hasConcept C2776214188 @default.
• W2799166683 hasConcept C2780451532 @default.
• W2799166683 hasConcept C41008148 @default.
• W2799166683 hasConcept C45374587 @default.
• W2799166683 hasConcept C55493867 @default.
• W2799166683 hasConcept C63479239 @default.
• W2799166683 hasConcept C6557445 @default.
• W2799166683 hasConcept C86803240 @default.
• W2799166683 hasConceptScore W2799166683C104317684 @default.
• W2799166683 hasConceptScore W2799166683C108010975 @default.
• W2799166683 hasConceptScore W2799166683C11413529 @default.
• W2799166683 hasConceptScore W2799166683C119857082 @default.
• W2799166683 hasConceptScore W2799166683C137293760 @default.
• W2799166683 hasConceptScore W2799166683C153083717 @default.
• W2799166683 hasConceptScore W2799166683C154945302 @default.
• W2799166683 hasConceptScore W2799166683C162324750 @default.
• W2799166683 hasConceptScore W2799166683C185592680 @default.
• W2799166683 hasConceptScore W2799166683C187736073 @default.
• W2799166683 hasConceptScore W2799166683C2776135515 @default.
• W2799166683 hasConceptScore W2799166683C2776214188 @default.
• W2799166683 hasConceptScore W2799166683C2780451532 @default.
• W2799166683 hasConceptScore W2799166683C41008148 @default.
• W2799166683 hasConceptScore W2799166683C45374587 @default.
• W2799166683 hasConceptScore W2799166683C55493867 @default.
• W2799166683 hasConceptScore W2799166683C63479239 @default.
• W2799166683 hasConceptScore W2799166683C6557445 @default.
• W2799166683 hasConceptScore W2799166683C86803240 @default.
• W2799166683 hasLocation W27991666831 @default.
• W2799166683 hasOpenAccess W2799166683 @default.
• W2799166683 hasPrimaryLocation W27991666831 @default.
• W2799166683 hasRelatedWork W2157560170 @default.
• W2799166683 hasRelatedWork W2356837744 @default.
• W2799166683 hasRelatedWork W2901381533 @default.
• W2799166683 hasRelatedWork W2914750564 @default.

• next