FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3119989665> ?p ?o ?g. }

• W3119989665 abstract "Pre-trained language models (LMs) are currently integral to many natural language processing systems. Although multilingual LMs were also introduced to serve many languages, these have limitations such as being costly at inference time and the size and diversity of non-English data involved in their pre-training. We remedy these issues for a collection of diverse Arabic varieties by introducing two powerful deep bidirectional transformer-based models, ARBERT and MARBERT. To evaluate our models, we also introduce ARLUE, a new benchmark for multi-dialectal Arabic language understanding evaluation. ARLUE is built using 42 datasets targeting six different task clusters, allowing us to offer a series of standardized experiments under rich conditions. When fine-tuned on ARLUE, our models collectively achieve new state-of-the-art results across the majority of tasks (37 out of 48 classification tasks, on the 42 datasets). Our best model acquires the highest ARLUE score (77.40) across all six task clusters, outperforming all other models including XLM-R Large (~ 3.4 x larger size). Our models are publicly available at this https URL and ARLUE will be released through the same repository." @default.
• W3119989665 created "2021-01-18" @default.
• W3119989665 creator A5004629670 @default.
• W3119989665 creator A5025365353 @default.
• W3119989665 creator A5041553790 @default.
• W3119989665 date "2020-12-27" @default.
• W3119989665 modified "2023-09-27" @default.
• W3119989665 title "ARBERT & MARBERT: Deep Bidirectional Transformers for Arabic." @default.
• W3119989665 cites W179850243 @default.
• W3119989665 cites W2027232045 @default.
• W3119989665 cites W2104912629 @default.
• W3119989665 cites W2121879602 @default.
• W3119989665 cites W2138923868 @default.
• W3119989665 cites W2160802179 @default.
• W3119989665 cites W2250594687 @default.
• W3119989665 cites W2252085349 @default.
• W3119989665 cites W2295710275 @default.
• W3119989665 cites W2396324390 @default.
• W3119989665 cites W2464521204 @default.
• W3119989665 cites W2525778437 @default.
• W3119989665 cites W2562238432 @default.
• W3119989665 cites W2621199241 @default.
• W3119989665 cites W2767566483 @default.
• W3119989665 cites W2770803436 @default.
• W3119989665 cites W2805351602 @default.
• W3119989665 cites W2805744755 @default.
• W3119989665 cites W2806092253 @default.
• W3119989665 cites W2809324505 @default.
• W3119989665 cites W2914220664 @default.
• W3119989665 cites W2916132663 @default.
• W3119989665 cites W2943552823 @default.
• W3119989665 cites W2945018970 @default.
• W3119989665 cites W2948433920 @default.
• W3119989665 cites W2948902769 @default.
• W3119989665 cites W2963310665 @default.
• W3119989665 cites W2963341956 @default.
• W3119989665 cites W2964121744 @default.
• W3119989665 cites W2965373594 @default.
• W3119989665 cites W2970485137 @default.
• W3119989665 cites W2970814728 @default.
• W3119989665 cites W2970960342 @default.
• W3119989665 cites W2972070042 @default.
• W3119989665 cites W2975059944 @default.
• W3119989665 cites W2995647371 @default.
• W3119989665 cites W2996580882 @default.
• W3119989665 cites W3000128329 @default.
• W3119989665 cites W3025939269 @default.
• W3119989665 cites W3032532958 @default.
• W3119989665 cites W3032746405 @default.
• W3119989665 cites W3033940819 @default.
• W3119989665 cites W3034469191 @default.
• W3119989665 cites W3035390927 @default.
• W3119989665 cites W3035497479 @default.
• W3119989665 cites W3040245432 @default.
• W3119989665 cites W3082274269 @default.
• W3119989665 cites W3087889451 @default.
• W3119989665 cites W3087891130 @default.
• W3119989665 cites W3088188607 @default.
• W3119989665 cites W3088592174 @default.
• W3119989665 cites W3088728183 @default.
• W3119989665 cites W3088774333 @default.
• W3119989665 cites W3088892776 @default.
• W3119989665 cites W3089190054 @default.
• W3119989665 cites W3098637735 @default.
• W3119989665 cites W3106433641 @default.
• W3119989665 cites W3118485687 @default.
• W3119989665 cites W3119566336 @default.
• W3119989665 cites W3119989085 @default.
• W3119989665 cites W3120373775 @default.
• W3119989665 cites W3154368324 @default.
• W3119989665 cites W3169483174 @default.
• W3119989665 cites W340195604 @default.
• W3119989665 hasPublicationYear "2020" @default.
• W3119989665 type Work @default.
• W3119989665 sameAs 3119989665 @default.
• W3119989665 citedByCount "33" @default.
• W3119989665 countsByYear W31199896652020 @default.
• W3119989665 countsByYear W31199896652021 @default.
• W3119989665 crossrefType "posted-content" @default.
• W3119989665 hasAuthorship W3119989665A5004629670 @default.
• W3119989665 hasAuthorship W3119989665A5025365353 @default.
• W3119989665 hasAuthorship W3119989665A5041553790 @default.
• W3119989665 hasConcept C119857082 @default.
• W3119989665 hasConcept C121332964 @default.
• W3119989665 hasConcept C13280743 @default.
• W3119989665 hasConcept C137293760 @default.
• W3119989665 hasConcept C138885662 @default.
• W3119989665 hasConcept C154945302 @default.
• W3119989665 hasConcept C162324750 @default.
• W3119989665 hasConcept C165801399 @default.
• W3119989665 hasConcept C185798385 @default.
• W3119989665 hasConcept C187736073 @default.
• W3119989665 hasConcept C204321447 @default.
• W3119989665 hasConcept C205649164 @default.
• W3119989665 hasConcept C2776214188 @default.
• W3119989665 hasConcept C2780451532 @default.
• W3119989665 hasConcept C41008148 @default.
• W3119989665 hasConcept C41895202 @default.
• W3119989665 hasConcept C62520636 @default.
• W3119989665 hasConcept C66322947 @default.

• next