Matches in SemOpenAlex for { <https://semopenalex.org/work/W3037474895> ?p ?o ?g. }
- W3037474895 abstract "This paper presents DeepSPIN’s submissions to Tasks 0 and 1 of the SIGMORPHON 2020 Shared Task. For both tasks, we present multilingual models, training jointly on data in all languages. We perform no language-specific hyperparameter tuning – each of our submissions uses the same model for all languages. Our basic architecture is the sparse sequence-to-sequence model with entmax attention and loss, which allows our models to learn sparse, local alignments while still being trainable with gradient-based techniques. For Task 1, we achieve strong performance with both RNN- and transformer-based sparse models. For Task 0, we extend our RNN-based model to a multi-encoder set-up in which separate modules encode the lemma and inflection sequences. Despite our models’ lack of language-specific tuning, they tie for first in Task 0 and place third in Task 1." @default.
- W3037474895 created "2020-07-02" @default.
- W3037474895 creator A5002087731 @default.
- W3037474895 creator A5051693368 @default.
- W3037474895 date "2020-01-01" @default.
- W3037474895 modified "2023-09-25" @default.
- W3037474895 title "One-Size-Fits-All Multilingual Models" @default.
- W3037474895 cites W1417002089 @default.
- W3037474895 cites W183625566 @default.
- W3037474895 cites W1902237438 @default.
- W3037474895 cites W1983874169 @default.
- W3037474895 cites W2064675550 @default.
- W3037474895 cites W2153579005 @default.
- W3037474895 cites W2187089797 @default.
- W3037474895 cites W2262393948 @default.
- W3037474895 cites W2351252181 @default.
- W3037474895 cites W2550821151 @default.
- W3037474895 cites W2566957588 @default.
- W3037474895 cites W2625720409 @default.
- W3037474895 cites W2781461271 @default.
- W3037474895 cites W2889368791 @default.
- W3037474895 cites W2899491443 @default.
- W3037474895 cites W2950825792 @default.
- W3037474895 cites W2950858167 @default.
- W3037474895 cites W2955719262 @default.
- W3037474895 cites W2962739339 @default.
- W3037474895 cites W2963123301 @default.
- W3037474895 cites W2963403868 @default.
- W3037474895 cites W2963407669 @default.
- W3037474895 cites W2963560568 @default.
- W3037474895 cites W2963635689 @default.
- W3037474895 cites W2963913768 @default.
- W3037474895 cites W2964034111 @default.
- W3037474895 cites W2964121744 @default.
- W3037474895 cites W2964308564 @default.
- W3037474895 cites W2970756316 @default.
- W3037474895 cites W2970963828 @default.
- W3037474895 cites W2972788883 @default.
- W3037474895 cites W2983833314 @default.
- W3037474895 cites W3010898928 @default.
- W3037474895 cites W3026538704 @default.
- W3037474895 cites W3037676302 @default.
- W3037474895 cites W3037995636 @default.
- W3037474895 cites W3023071679 @default.
- W3037474895 doi "https://doi.org/10.18653/v1/2020.sigmorphon-1.4" @default.
- W3037474895 hasPublicationYear "2020" @default.
- W3037474895 type Work @default.
- W3037474895 sameAs 3037474895 @default.
- W3037474895 citedByCount "5" @default.
- W3037474895 countsByYear W30374748952020 @default.
- W3037474895 countsByYear W30374748952021 @default.
- W3037474895 countsByYear W30374748952022 @default.
- W3037474895 crossrefType "proceedings-article" @default.
- W3037474895 hasAuthorship W3037474895A5002087731 @default.
- W3037474895 hasAuthorship W3037474895A5051693368 @default.
- W3037474895 hasBestOaLocation W30374748951 @default.
- W3037474895 hasConcept C104317684 @default.
- W3037474895 hasConcept C111919701 @default.
- W3037474895 hasConcept C118505674 @default.
- W3037474895 hasConcept C119857082 @default.
- W3037474895 hasConcept C121332964 @default.
- W3037474895 hasConcept C137293760 @default.
- W3037474895 hasConcept C154945302 @default.
- W3037474895 hasConcept C159403335 @default.
- W3037474895 hasConcept C162324750 @default.
- W3037474895 hasConcept C165801399 @default.
- W3037474895 hasConcept C177264268 @default.
- W3037474895 hasConcept C185592680 @default.
- W3037474895 hasConcept C187736073 @default.
- W3037474895 hasConcept C18903297 @default.
- W3037474895 hasConcept C199360897 @default.
- W3037474895 hasConcept C204321447 @default.
- W3037474895 hasConcept C2777759810 @default.
- W3037474895 hasConcept C2780451532 @default.
- W3037474895 hasConcept C28006648 @default.
- W3037474895 hasConcept C34736171 @default.
- W3037474895 hasConcept C41008148 @default.
- W3037474895 hasConcept C46757340 @default.
- W3037474895 hasConcept C55493867 @default.
- W3037474895 hasConcept C62520636 @default.
- W3037474895 hasConcept C66322947 @default.
- W3037474895 hasConcept C66746571 @default.
- W3037474895 hasConcept C8642999 @default.
- W3037474895 hasConcept C86803240 @default.
- W3037474895 hasConceptScore W3037474895C104317684 @default.
- W3037474895 hasConceptScore W3037474895C111919701 @default.
- W3037474895 hasConceptScore W3037474895C118505674 @default.
- W3037474895 hasConceptScore W3037474895C119857082 @default.
- W3037474895 hasConceptScore W3037474895C121332964 @default.
- W3037474895 hasConceptScore W3037474895C137293760 @default.
- W3037474895 hasConceptScore W3037474895C154945302 @default.
- W3037474895 hasConceptScore W3037474895C159403335 @default.
- W3037474895 hasConceptScore W3037474895C162324750 @default.
- W3037474895 hasConceptScore W3037474895C165801399 @default.
- W3037474895 hasConceptScore W3037474895C177264268 @default.
- W3037474895 hasConceptScore W3037474895C185592680 @default.
- W3037474895 hasConceptScore W3037474895C187736073 @default.
- W3037474895 hasConceptScore W3037474895C18903297 @default.
- W3037474895 hasConceptScore W3037474895C199360897 @default.
- W3037474895 hasConceptScore W3037474895C204321447 @default.