FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3100268441> ?p ?o ?g. }

• W3100268441 abstract "Despite their empirical success, neural networks still have difficulty capturing compositional aspects of natural language. This work proposes a simple data augmentation approach to encourage compositional behavior in neural models for sequence-to-sequence problems. Our approach, SeqMix, creates new synthetic examples by softly combining input/output sequences from the training set. We connect this approach to existing techniques such as SwitchOut and word dropout, and show that these techniques are all essentially approximating variants of a single objective. SeqMix consistently yields approximately 1.0 BLEU improvement on five different translation datasets over strong Transformer baselines. On tasks that require strong compositional generalization such as SCAN and semantic parsing, SeqMix also offers further improvements." @default.
• W3100268441 created "2020-11-23" @default.
• W3100268441 creator A5045913119 @default.
• W3100268441 creator A5049260750 @default.
• W3100268441 creator A5086644493 @default.
• W3100268441 date "2020-01-01" @default.
• W3100268441 modified "2023-10-16" @default.
• W3100268441 title "Sequence-Level Mixed Sample Data Augmentation" @default.
• W3100268441 cites W1514535095 @default.
• W3100268441 cites W2118373646 @default.
• W3100268441 cites W2251939518 @default.
• W3100268441 cites W2308720496 @default.
• W3100268441 cites W2746314669 @default.
• W3100268441 cites W2798663534 @default.
• W3100268441 cites W2859444450 @default.
• W3100268441 cites W2888519496 @default.
• W3100268441 cites W2889368791 @default.
• W3100268441 cites W2890909908 @default.
• W3100268441 cites W2921861056 @default.
• W3100268441 cites W2933138175 @default.
• W3100268441 cites W2953173959 @default.
• W3100268441 cites W2962716332 @default.
• W3100268441 cites W2962801832 @default.
• W3100268441 cites W2963073938 @default.
• W3100268441 cites W2963267799 @default.
• W3100268441 cites W2963399829 @default.
• W3100268441 cites W2963863610 @default.
• W3100268441 cites W2964118342 @default.
• W3100268441 cites W2992308087 @default.
• W3100268441 cites W3035331128 @default.
• W3100268441 cites W3035542229 @default.
• W3100268441 doi "https://doi.org/10.18653/v1/2020.emnlp-main.447" @default.
• W3100268441 hasPublicationYear "2020" @default.
• W3100268441 type Work @default.
• W3100268441 sameAs 3100268441 @default.
• W3100268441 citedByCount "31" @default.
• W3100268441 countsByYear W31002684412021 @default.
• W3100268441 countsByYear W31002684412022 @default.
• W3100268441 countsByYear W31002684412023 @default.
• W3100268441 crossrefType "proceedings-article" @default.
• W3100268441 hasAuthorship W3100268441A5045913119 @default.
• W3100268441 hasAuthorship W3100268441A5049260750 @default.
• W3100268441 hasAuthorship W3100268441A5086644493 @default.
• W3100268441 hasBestOaLocation W31002684411 @default.
• W3100268441 hasConcept C119857082 @default.
• W3100268441 hasConcept C121332964 @default.
• W3100268441 hasConcept C134306372 @default.
• W3100268441 hasConcept C154945302 @default.
• W3100268441 hasConcept C165801399 @default.
• W3100268441 hasConcept C177148314 @default.
• W3100268441 hasConcept C177264268 @default.
• W3100268441 hasConcept C185592680 @default.
• W3100268441 hasConcept C186644900 @default.
• W3100268441 hasConcept C195324797 @default.
• W3100268441 hasConcept C198531522 @default.
• W3100268441 hasConcept C199360897 @default.
• W3100268441 hasConcept C203005215 @default.
• W3100268441 hasConcept C204321447 @default.
• W3100268441 hasConcept C2524010 @default.
• W3100268441 hasConcept C2776145597 @default.
• W3100268441 hasConcept C2778112365 @default.
• W3100268441 hasConcept C33923547 @default.
• W3100268441 hasConcept C41008148 @default.
• W3100268441 hasConcept C43617362 @default.
• W3100268441 hasConcept C50644808 @default.
• W3100268441 hasConcept C51632099 @default.
• W3100268441 hasConcept C54355233 @default.
• W3100268441 hasConcept C62520636 @default.
• W3100268441 hasConcept C66322947 @default.
• W3100268441 hasConcept C86803240 @default.
• W3100268441 hasConcept C90805587 @default.
• W3100268441 hasConceptScore W3100268441C119857082 @default.
• W3100268441 hasConceptScore W3100268441C121332964 @default.
• W3100268441 hasConceptScore W3100268441C134306372 @default.
• W3100268441 hasConceptScore W3100268441C154945302 @default.
• W3100268441 hasConceptScore W3100268441C165801399 @default.
• W3100268441 hasConceptScore W3100268441C177148314 @default.
• W3100268441 hasConceptScore W3100268441C177264268 @default.
• W3100268441 hasConceptScore W3100268441C185592680 @default.
• W3100268441 hasConceptScore W3100268441C186644900 @default.
• W3100268441 hasConceptScore W3100268441C195324797 @default.
• W3100268441 hasConceptScore W3100268441C198531522 @default.
• W3100268441 hasConceptScore W3100268441C199360897 @default.
• W3100268441 hasConceptScore W3100268441C203005215 @default.
• W3100268441 hasConceptScore W3100268441C204321447 @default.
• W3100268441 hasConceptScore W3100268441C2524010 @default.
• W3100268441 hasConceptScore W3100268441C2776145597 @default.
• W3100268441 hasConceptScore W3100268441C2778112365 @default.
• W3100268441 hasConceptScore W3100268441C33923547 @default.
• W3100268441 hasConceptScore W3100268441C41008148 @default.
• W3100268441 hasConceptScore W3100268441C43617362 @default.
• W3100268441 hasConceptScore W3100268441C50644808 @default.
• W3100268441 hasConceptScore W3100268441C51632099 @default.
• W3100268441 hasConceptScore W3100268441C54355233 @default.
• W3100268441 hasConceptScore W3100268441C62520636 @default.
• W3100268441 hasConceptScore W3100268441C66322947 @default.
• W3100268441 hasConceptScore W3100268441C86803240 @default.
• W3100268441 hasConceptScore W3100268441C90805587 @default.
• W3100268441 hasLocation W31002684411 @default.
• W3100268441 hasLocation W31002684412 @default.

• next