FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3096565276> ?p ?o ?g. }

• W3096565276 abstract "State-of-the-art natural language understanding classification models follow two-stages: pre-training a large language model on an auxiliary task, and then fine-tuning the model on a task-specific labeled dataset using cross-entropy loss. However, the cross-entropy loss has several shortcomings that can lead to sub-optimal generalization and instability. Driven by the intuition that good generalization requires capturing the similarity between examples in one class and contrasting them with examples in other classes, we propose a supervised contrastive learning (SCL) objective for the fine-tuning stage. Combined with cross-entropy, our proposed SCL loss obtains significant improvements over a strong RoBERTa-Large baseline on multiple datasets of the GLUE benchmark in few-shot learning settings, without requiring specialized architecture, data augmentations, memory banks, or additional unsupervised data. Our proposed fine-tuning objective leads to models that are more robust to different levels of noise in the fine-tuning training data, and can generalize better to related tasks with limited labeled data." @default.
• W3096565276 created "2020-11-09" @default.
• W3096565276 creator A5005030122 @default.
• W3096565276 creator A5006191011 @default.
• W3096565276 creator A5068394403 @default.
• W3096565276 creator A5091071525 @default.
• W3096565276 date "2020-11-03" @default.
• W3096565276 modified "2023-10-03" @default.
• W3096565276 title "Supervised Contrastive Learning for Pre-trained Language Model Fine-tuning" @default.
• W3096565276 cites W1544827683 @default.
• W3096565276 cites W1821462560 @default.
• W3096565276 cites W1840435438 @default.
• W3096565276 cites W2097732278 @default.
• W3096565276 cites W2117539524 @default.
• W3096565276 cites W2138204974 @default.
• W3096565276 cites W2163614729 @default.
• W3096565276 cites W2171849160 @default.
• W3096565276 cites W2183341477 @default.
• W3096565276 cites W2194775991 @default.
• W3096565276 cites W2293778248 @default.
• W3096565276 cites W2555897561 @default.
• W3096565276 cites W2788496822 @default.
• W3096565276 cites W2798991696 @default.
• W3096565276 cites W2842511635 @default.
• W3096565276 cites W2887997457 @default.
• W3096565276 cites W2889326796 @default.
• W3096565276 cites W2913881544 @default.
• W3096565276 cites W2913939497 @default.
• W3096565276 cites W2917551568 @default.
• W3096565276 cites W2933138175 @default.
• W3096565276 cites W2943152387 @default.
• W3096565276 cites W2944828972 @default.
• W3096565276 cites W2949736877 @default.
• W3096565276 cites W2962369866 @default.
• W3096565276 cites W2962788840 @default.
• W3096565276 cites W2963012544 @default.
• W3096565276 cites W2963026768 @default.
• W3096565276 cites W2963060032 @default.
• W3096565276 cites W2963310665 @default.
• W3096565276 cites W2963341956 @default.
• W3096565276 cites W2963399829 @default.
• W3096565276 cites W2963656735 @default.
• W3096565276 cites W2963759070 @default.
• W3096565276 cites W2964013229 @default.
• W3096565276 cites W2965373594 @default.
• W3096565276 cites W2970206392 @default.
• W3096565276 cites W2970941190 @default.
• W3096565276 cites W2971155163 @default.
• W3096565276 cites W2973562770 @default.
• W3096565276 cites W2992308087 @default.
• W3096565276 cites W3005700362 @default.
• W3096565276 cites W3010874390 @default.
• W3096565276 cites W3026732421 @default.
• W3096565276 cites W3034709122 @default.
• W3096565276 cites W3034781633 @default.
• W3096565276 cites W3034850762 @default.
• W3096565276 cites W3034978746 @default.
• W3096565276 cites W3035160371 @default.
• W3096565276 cites W3035204084 @default.
• W3096565276 cites W3035524453 @default.
• W3096565276 cites W3037057938 @default.
• W3096565276 cites W3043462782 @default.
• W3096565276 cites W3099782249 @default.
• W3096565276 cites W3100345210 @default.
• W3096565276 cites W3100859887 @default.
• W3096565276 cites W3102631365 @default.
• W3096565276 cites W3108655343 @default.
• W3096565276 cites W3118608800 @default.
• W3096565276 hasPublicationYear "2020" @default.
• W3096565276 type Work @default.
• W3096565276 sameAs 3096565276 @default.
• W3096565276 citedByCount "9" @default.
• W3096565276 countsByYear W30965652762021 @default.
• W3096565276 crossrefType "posted-content" @default.
• W3096565276 hasAuthorship W3096565276A5005030122 @default.
• W3096565276 hasAuthorship W3096565276A5006191011 @default.
• W3096565276 hasAuthorship W3096565276A5068394403 @default.
• W3096565276 hasAuthorship W3096565276A5091071525 @default.
• W3096565276 hasConcept C106301342 @default.
• W3096565276 hasConcept C119857082 @default.
• W3096565276 hasConcept C121332964 @default.
• W3096565276 hasConcept C134306372 @default.
• W3096565276 hasConcept C153180895 @default.
• W3096565276 hasConcept C154945302 @default.
• W3096565276 hasConcept C157524613 @default.
• W3096565276 hasConcept C167981619 @default.
• W3096565276 hasConcept C177148314 @default.
• W3096565276 hasConcept C204321447 @default.
• W3096565276 hasConcept C33923547 @default.
• W3096565276 hasConcept C41008148 @default.
• W3096565276 hasConcept C62520636 @default.
• W3096565276 hasConceptScore W3096565276C106301342 @default.
• W3096565276 hasConceptScore W3096565276C119857082 @default.
• W3096565276 hasConceptScore W3096565276C121332964 @default.
• W3096565276 hasConceptScore W3096565276C134306372 @default.
• W3096565276 hasConceptScore W3096565276C153180895 @default.
• W3096565276 hasConceptScore W3096565276C154945302 @default.
• W3096565276 hasConceptScore W3096565276C157524613 @default.
• W3096565276 hasConceptScore W3096565276C167981619 @default.
• W3096565276 hasConceptScore W3096565276C177148314 @default.

• next