FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3199553313> ?p ?o ?g. }

• W3199553313 abstract "Knowledge Distillation (KD) is extensively used to compress and deploy large pre-trained language models on edge devices for real-world applications. However, one neglected area of research is the impact of noisy (corrupted) labels on KD. We present, to the best of our knowledge, the first study on KD with noisy labels in Natural Language Understanding (NLU). We document the scope of the problem and present two methods to mitigate the impact of label noise. Experiments on the GLUE benchmark show that our methods are effective even under high noise levels. Nevertheless, our results indicate that more research is necessary to cope with label noise under the KD." @default.
• W3199553313 created "2021-09-27" @default.
• W3199553313 creator A5004641996 @default.
• W3199553313 creator A5019296120 @default.
• W3199553313 creator A5022989187 @default.
• W3199553313 creator A5028862918 @default.
• W3199553313 creator A5034154991 @default.
• W3199553313 creator A5040035859 @default.
• W3199553313 creator A5044793533 @default.
• W3199553313 creator A5066020726 @default.
• W3199553313 date "2021-01-01" @default.
• W3199553313 modified "2023-10-14" @default.
• W3199553313 title "Knowledge Distillation with Noisy Labels for Natural Language Understanding" @default.
• W3199553313 cites W1821462560 @default.
• W3199553313 cites W2101234009 @default.
• W3199553313 cites W2123838014 @default.
• W3199553313 cites W2167460663 @default.
• W3199553313 cites W2294370754 @default.
• W3199553313 cites W2743200750 @default.
• W3199553313 cites W2769240247 @default.
• W3199553313 cites W2802198257 @default.
• W3199553313 cites W2896457183 @default.
• W3199553313 cites W2911964244 @default.
• W3199553313 cites W2941387380 @default.
• W3199553313 cites W2946376679 @default.
• W3199553313 cites W2963310665 @default.
• W3199553313 cites W2963323070 @default.
• W3199553313 cites W2963371670 @default.
• W3199553313 cites W2963386594 @default.
• W3199553313 cites W2963735582 @default.
• W3199553313 cites W2963999980 @default.
• W3199553313 cites W2975429091 @default.
• W3199553313 cites W2978017171 @default.
• W3199553313 cites W2978544343 @default.
• W3199553313 cites W2978625989 @default.
• W3199553313 cites W2981873476 @default.
• W3199553313 cites W2985678088 @default.
• W3199553313 cites W2990704537 @default.
• W3199553313 cites W2997941347 @default.
• W3199553313 cites W2998239226 @default.
• W3199553313 cites W3008374555 @default.
• W3199553313 cites W3030163527 @default.
• W3199553313 cites W3034457371 @default.
• W3199553313 cites W3041342983 @default.
• W3199553313 cites W3082274269 @default.
• W3199553313 cites W3092173171 @default.
• W3199553313 cites W3103846556 @default.
• W3199553313 cites W3105966348 @default.
• W3199553313 cites W3120832022 @default.
• W3199553313 cites W3121951843 @default.
• W3199553313 cites W3152607317 @default.
• W3199553313 cites W3174731106 @default.
• W3199553313 cites W3175002116 @default.
• W3199553313 cites W3175234294 @default.
• W3199553313 cites W3175352680 @default.
• W3199553313 cites W3177378457 @default.
• W3199553313 cites W3186056917 @default.
• W3199553313 cites W3213820586 @default.
• W3199553313 doi "https://doi.org/10.18653/v1/2021.wnut-1.33" @default.
• W3199553313 hasPublicationYear "2021" @default.
• W3199553313 type Work @default.
• W3199553313 sameAs 3199553313 @default.
• W3199553313 citedByCount "0" @default.
• W3199553313 crossrefType "proceedings-article" @default.
• W3199553313 hasAuthorship W3199553313A5004641996 @default.
• W3199553313 hasAuthorship W3199553313A5019296120 @default.
• W3199553313 hasAuthorship W3199553313A5022989187 @default.
• W3199553313 hasAuthorship W3199553313A5028862918 @default.
• W3199553313 hasAuthorship W3199553313A5034154991 @default.
• W3199553313 hasAuthorship W3199553313A5040035859 @default.
• W3199553313 hasAuthorship W3199553313A5044793533 @default.
• W3199553313 hasAuthorship W3199553313A5066020726 @default.
• W3199553313 hasBestOaLocation W31995533131 @default.
• W3199553313 hasConcept C115961682 @default.
• W3199553313 hasConcept C119857082 @default.
• W3199553313 hasConcept C13280743 @default.
• W3199553313 hasConcept C154945302 @default.
• W3199553313 hasConcept C162307627 @default.
• W3199553313 hasConcept C166957645 @default.
• W3199553313 hasConcept C178790620 @default.
• W3199553313 hasConcept C185592680 @default.
• W3199553313 hasConcept C185798385 @default.
• W3199553313 hasConcept C195324797 @default.
• W3199553313 hasConcept C199360897 @default.
• W3199553313 hasConcept C204030448 @default.
• W3199553313 hasConcept C204321447 @default.
• W3199553313 hasConcept C205649164 @default.
• W3199553313 hasConcept C2776608160 @default.
• W3199553313 hasConcept C2778012447 @default.
• W3199553313 hasConcept C2779439875 @default.
• W3199553313 hasConcept C41008148 @default.
• W3199553313 hasConcept C95457728 @default.
• W3199553313 hasConcept C99498987 @default.
• W3199553313 hasConceptScore W3199553313C115961682 @default.
• W3199553313 hasConceptScore W3199553313C119857082 @default.
• W3199553313 hasConceptScore W3199553313C13280743 @default.
• W3199553313 hasConceptScore W3199553313C154945302 @default.
• W3199553313 hasConceptScore W3199553313C162307627 @default.
• W3199553313 hasConceptScore W3199553313C166957645 @default.
• W3199553313 hasConceptScore W3199553313C178790620 @default.

• next