FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3199212312> ?p ?o ?g. }

• W3199212312 abstract "We explore the impact of leveraging the relatedness of languages that belong to the same family in NLP models using multilingual fine-tuning. We hypothesize and validate that multilingual fine-tuning of pre-trained language models can yield better performance on downstream NLP applications, compared to models fine-tuned on individual languages. A first of its kind detailed study is presented to track performance change as languages are added to a base language in a graded and greedy (in the sense of best boost of performance) manner; which reveals that careful selection of subset of related languages can significantly improve performance than utilizing all related languages. The Indo-Aryan (IA) language family is chosen for the study, the exact languages being Bengali, Gujarati, Hindi, Marathi, Oriya, Punjabi and Urdu. The script barrier is crossed by simple rule-based transliteration of the text of all languages to Devanagari. Experiments are performed on mBERT, IndicBERT, MuRIL and two RoBERTa-based LMs, the last two being pre-trained by us. Low resource languages, such as Oriya and Punjabi, are found to be the largest beneficiaries of multilingual fine-tuning. Textual Entailment, Entity Classification, Section Title Prediction, tasks of IndicGLUE and POS tagging form our test bed. Compared to monolingual fine tuning we get relative performance improvement of up to 150% in the downstream tasks. The surprise take-away is that for any language there is a particular combination of other languages which yields the best performance, and any additional language is in fact detrimental." @default.
• W3199212312 created "2021-09-27" @default.
• W3199212312 creator A5000493089 @default.
• W3199212312 creator A5004454662 @default.
• W3199212312 creator A5017064237 @default.
• W3199212312 creator A5060561123 @default.
• W3199212312 creator A5065100828 @default.
• W3199212312 date "2021-01-01" @default.
• W3199212312 modified "2023-10-16" @default.
• W3199212312 title "Role of Language Relatedness in Multilingual Fine-tuning of Language Models: A Case Study in Indo-Aryan Languages" @default.
• W3199212312 cites W2251467004 @default.
• W3199212312 cites W2273712250 @default.
• W3199212312 cites W2278252977 @default.
• W3199212312 cites W2465297109 @default.
• W3199212312 cites W2794998650 @default.
• W3199212312 cites W2914120296 @default.
• W3199212312 cites W2963341956 @default.
• W3199212312 cites W2963403868 @default.
• W3199212312 cites W2965373594 @default.
• W3199212312 cites W2970557265 @default.
• W3199212312 cites W2970597249 @default.
• W3199212312 cites W2970854433 @default.
• W3199212312 cites W2972688845 @default.
• W3199212312 cites W2982399380 @default.
• W3199212312 cites W2988257285 @default.
• W3199212312 cites W2994885767 @default.
• W3199212312 cites W2998353611 @default.
• W3199212312 cites W3028872947 @default.
• W3199212312 cites W3035390927 @default.
• W3199212312 cites W3035547806 @default.
• W3199212312 cites W3097821138 @default.
• W3199212312 cites W3098824823 @default.
• W3199212312 cites W3099919888 @default.
• W3199212312 cites W3102425047 @default.
• W3199212312 cites W3105005398 @default.
• W3199212312 cites W3107826490 @default.
• W3199212312 cites W3161740710 @default.
• W3199212312 doi "https://doi.org/10.18653/v1/2021.emnlp-main.675" @default.
• W3199212312 hasPublicationYear "2021" @default.
• W3199212312 type Work @default.
• W3199212312 sameAs 3199212312 @default.
• W3199212312 citedByCount "2" @default.
• W3199212312 countsByYear W31992123122023 @default.
• W3199212312 crossrefType "proceedings-article" @default.
• W3199212312 hasAuthorship W3199212312A5000493089 @default.
• W3199212312 hasAuthorship W3199212312A5004454662 @default.
• W3199212312 hasAuthorship W3199212312A5017064237 @default.
• W3199212312 hasAuthorship W3199212312A5060561123 @default.
• W3199212312 hasAuthorship W3199212312A5065100828 @default.
• W3199212312 hasBestOaLocation W31992123121 @default.
• W3199212312 hasConcept C115961682 @default.
• W3199212312 hasConcept C137293760 @default.
• W3199212312 hasConcept C138885662 @default.
• W3199212312 hasConcept C154945302 @default.
• W3199212312 hasConcept C19235068 @default.
• W3199212312 hasConcept C204321447 @default.
• W3199212312 hasConcept C2776844415 @default.
• W3199212312 hasConcept C2777350258 @default.
• W3199212312 hasConcept C2778756302 @default.
• W3199212312 hasConcept C2779662586 @default.
• W3199212312 hasConcept C2780144916 @default.
• W3199212312 hasConcept C2987247673 @default.
• W3199212312 hasConcept C41008148 @default.
• W3199212312 hasConcept C41895202 @default.
• W3199212312 hasConcept C519982507 @default.
• W3199212312 hasConcept C520968082 @default.
• W3199212312 hasConcept C81917197 @default.
• W3199212312 hasConceptScore W3199212312C115961682 @default.
• W3199212312 hasConceptScore W3199212312C137293760 @default.
• W3199212312 hasConceptScore W3199212312C138885662 @default.
• W3199212312 hasConceptScore W3199212312C154945302 @default.
• W3199212312 hasConceptScore W3199212312C19235068 @default.
• W3199212312 hasConceptScore W3199212312C204321447 @default.
• W3199212312 hasConceptScore W3199212312C2776844415 @default.
• W3199212312 hasConceptScore W3199212312C2777350258 @default.
• W3199212312 hasConceptScore W3199212312C2778756302 @default.
• W3199212312 hasConceptScore W3199212312C2779662586 @default.
• W3199212312 hasConceptScore W3199212312C2780144916 @default.
• W3199212312 hasConceptScore W3199212312C2987247673 @default.
• W3199212312 hasConceptScore W3199212312C41008148 @default.
• W3199212312 hasConceptScore W3199212312C41895202 @default.
• W3199212312 hasConceptScore W3199212312C519982507 @default.
• W3199212312 hasConceptScore W3199212312C520968082 @default.
• W3199212312 hasConceptScore W3199212312C81917197 @default.
• W3199212312 hasLocation W31992123121 @default.
• W3199212312 hasLocation W31992123122 @default.
• W3199212312 hasOpenAccess W3199212312 @default.
• W3199212312 hasPrimaryLocation W31992123121 @default.
• W3199212312 hasRelatedWork W170711724 @default.
• W3199212312 hasRelatedWork W2316281861 @default.
• W3199212312 hasRelatedWork W2562065833 @default.
• W3199212312 hasRelatedWork W3031586918 @default.
• W3199212312 hasRelatedWork W3043528814 @default.
• W3199212312 hasRelatedWork W3184477063 @default.
• W3199212312 hasRelatedWork W3199212312 @default.
• W3199212312 hasRelatedWork W3211583198 @default.
• W3199212312 hasRelatedWork W4286967713 @default.
• W3199212312 hasRelatedWork W4297712837 @default.
• W3199212312 isParatext "false" @default.
• W3199212312 isRetracted "false" @default.

• next