Matches in SemOpenAlex for { <https://semopenalex.org/work/W3089477688> ?p ?o ?g. }
Showing items 1 to 91 of
91
with 100 items per page.
- W3089477688 abstract "It is challenging to perform lifelong language learning (LLL) on a stream of different tasks without any performance degradation comparing to the multi-task counterparts. To address this issue, we present Lifelong Language Knowledge Distillation (L2KD), a simple but efficient method that can be easily applied to existing LLL architectures in order to mitigate the degradation. Specifically, when the LLL model is trained on a new task, we assign a teacher model to first learn the new task, and pass the knowledge to the LLL model via knowledge distillation. Therefore, the LLL model can better adapt to the new task while keeping the previously learned knowledge. Experiments show that the proposed L2KD consistently improves previous state-of-the-art models, and the degradation comparing to multi-task models in LLL tasks is well mitigated for both sequence generation and text classification tasks." @default.
- W3089477688 created "2020-10-08" @default.
- W3089477688 creator A5058729228 @default.
- W3089477688 creator A5076610826 @default.
- W3089477688 creator A5077356556 @default.
- W3089477688 date "2020-10-05" @default.
- W3089477688 modified "2023-10-18" @default.
- W3089477688 title "Lifelong Language Knowledge Distillation" @default.
- W3089477688 cites W1682403713 @default.
- W3089477688 cites W1821462560 @default.
- W3089477688 cites W2294370754 @default.
- W3089477688 cites W2296073425 @default.
- W3089477688 cites W2606974598 @default.
- W3089477688 cites W2734314755 @default.
- W3089477688 cites W2751448157 @default.
- W3089477688 cites W2767206889 @default.
- W3089477688 cites W2767245334 @default.
- W3089477688 cites W2777054756 @default.
- W3089477688 cites W2809324505 @default.
- W3089477688 cites W2894094671 @default.
- W3089477688 cites W2895723011 @default.
- W3089477688 cites W2936858556 @default.
- W3089477688 cites W2948743095 @default.
- W3089477688 cites W2951422922 @default.
- W3089477688 cites W2952013107 @default.
- W3089477688 cites W2962860923 @default.
- W3089477688 cites W2963453233 @default.
- W3089477688 cites W2963588172 @default.
- W3089477688 cites W2964222566 @default.
- W3089477688 cites W3002800333 @default.
- W3089477688 doi "https://doi.org/10.48550/arxiv.2010.02123" @default.
- W3089477688 hasPublicationYear "2020" @default.
- W3089477688 type Work @default.
- W3089477688 sameAs 3089477688 @default.
- W3089477688 citedByCount "5" @default.
- W3089477688 countsByYear W30894776882020 @default.
- W3089477688 countsByYear W30894776882021 @default.
- W3089477688 crossrefType "posted-content" @default.
- W3089477688 hasAuthorship W3089477688A5058729228 @default.
- W3089477688 hasAuthorship W3089477688A5076610826 @default.
- W3089477688 hasAuthorship W3089477688A5077356556 @default.
- W3089477688 hasBestOaLocation W30894776881 @default.
- W3089477688 hasConcept C108771440 @default.
- W3089477688 hasConcept C119857082 @default.
- W3089477688 hasConcept C127413603 @default.
- W3089477688 hasConcept C137293760 @default.
- W3089477688 hasConcept C154945302 @default.
- W3089477688 hasConcept C15744967 @default.
- W3089477688 hasConcept C178790620 @default.
- W3089477688 hasConcept C185592680 @default.
- W3089477688 hasConcept C19417346 @default.
- W3089477688 hasConcept C201995342 @default.
- W3089477688 hasConcept C204030448 @default.
- W3089477688 hasConcept C204321447 @default.
- W3089477688 hasConcept C2779679103 @default.
- W3089477688 hasConcept C2780451532 @default.
- W3089477688 hasConcept C41008148 @default.
- W3089477688 hasConcept C76155785 @default.
- W3089477688 hasConceptScore W3089477688C108771440 @default.
- W3089477688 hasConceptScore W3089477688C119857082 @default.
- W3089477688 hasConceptScore W3089477688C127413603 @default.
- W3089477688 hasConceptScore W3089477688C137293760 @default.
- W3089477688 hasConceptScore W3089477688C154945302 @default.
- W3089477688 hasConceptScore W3089477688C15744967 @default.
- W3089477688 hasConceptScore W3089477688C178790620 @default.
- W3089477688 hasConceptScore W3089477688C185592680 @default.
- W3089477688 hasConceptScore W3089477688C19417346 @default.
- W3089477688 hasConceptScore W3089477688C201995342 @default.
- W3089477688 hasConceptScore W3089477688C204030448 @default.
- W3089477688 hasConceptScore W3089477688C204321447 @default.
- W3089477688 hasConceptScore W3089477688C2779679103 @default.
- W3089477688 hasConceptScore W3089477688C2780451532 @default.
- W3089477688 hasConceptScore W3089477688C41008148 @default.
- W3089477688 hasConceptScore W3089477688C76155785 @default.
- W3089477688 hasLocation W30894776881 @default.
- W3089477688 hasOpenAccess W3089477688 @default.
- W3089477688 hasPrimaryLocation W30894776881 @default.
- W3089477688 hasRelatedWork W10596858 @default.
- W3089477688 hasRelatedWork W11991885 @default.
- W3089477688 hasRelatedWork W14230040 @default.
- W3089477688 hasRelatedWork W149980 @default.
- W3089477688 hasRelatedWork W6633147 @default.
- W3089477688 hasRelatedWork W7401400 @default.
- W3089477688 hasRelatedWork W9218159 @default.
- W3089477688 hasRelatedWork W9280962 @default.
- W3089477688 hasRelatedWork W1226999 @default.
- W3089477688 hasRelatedWork W13002482 @default.
- W3089477688 isParatext "false" @default.
- W3089477688 isRetracted "false" @default.
- W3089477688 magId "3089477688" @default.
- W3089477688 workType "article" @default.