Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385572058> ?p ?o ?g. }
Showing items 1 to 63 of
63
with 100 items per page.
- W4385572058 abstract "Adapters have emerged as a modular and parameter-efficient approach to (zero-shot) cross-lingual transfer. The established MAD-X framework employs separate language and task adapters which can be arbitrarily combined to perform the transfer of any task to any target language. Subsequently, BAD-X, an extension of the MAD-X framework, achieves improved transfer at the cost of MAD-X’s modularity by creating ‘bilingual’ adapters specific to the source-target language pair. In this work, we aim to take the best of both worlds by (i) fine-tuning *task* adapters adapted to the target language(s) (so-called *‘target language-ready’ (TLR)* adapters) to maintain high transfer performance, but (ii) without sacrificing the highly modular design of MAD-X. The main idea of ‘target language-ready’ adapters is to resolve the training-vs-inference discrepancy of MAD-X: the task adapter ‘sees’ the target language adapter for the very first time during inference, and thus might not be fully compatible with it. We address this mismatch by exposing the task adapter to the target language adapter during training, and empirically validate several variants of the idea: in the simplest form, we alternate between using the source and target language adapters during task adapter training, which can be generalized to cycling over any set of language adapters. We evaluate different TLR-based transfer configurations with varying degrees of generality across a suite of standard cross-lingual benchmarks, and find that the most general (and thus most modular) configuration consistently outperforms MAD-X and BAD-X on most tasks and languages." @default.
- W4385572058 created "2023-08-05" @default.
- W4385572058 creator A5014866912 @default.
- W4385572058 creator A5034773582 @default.
- W4385572058 creator A5056466747 @default.
- W4385572058 creator A5081393566 @default.
- W4385572058 date "2023-01-01" @default.
- W4385572058 modified "2023-09-27" @default.
- W4385572058 title "Cross-Lingual Transfer with Target Language-Ready Task Adapters" @default.
- W4385572058 doi "https://doi.org/10.18653/v1/2023.findings-acl.13" @default.
- W4385572058 hasPublicationYear "2023" @default.
- W4385572058 type Work @default.
- W4385572058 citedByCount "0" @default.
- W4385572058 crossrefType "proceedings-article" @default.
- W4385572058 hasAuthorship W4385572058A5014866912 @default.
- W4385572058 hasAuthorship W4385572058A5034773582 @default.
- W4385572058 hasAuthorship W4385572058A5056466747 @default.
- W4385572058 hasAuthorship W4385572058A5081393566 @default.
- W4385572058 hasBestOaLocation W43855720581 @default.
- W4385572058 hasConcept C101468663 @default.
- W4385572058 hasConcept C127413603 @default.
- W4385572058 hasConcept C137293760 @default.
- W4385572058 hasConcept C154945302 @default.
- W4385572058 hasConcept C173608175 @default.
- W4385572058 hasConcept C177284502 @default.
- W4385572058 hasConcept C199360897 @default.
- W4385572058 hasConcept C201995342 @default.
- W4385572058 hasConcept C204321447 @default.
- W4385572058 hasConcept C2776175482 @default.
- W4385572058 hasConcept C2776214188 @default.
- W4385572058 hasConcept C2780451532 @default.
- W4385572058 hasConcept C41008148 @default.
- W4385572058 hasConcept C9390403 @default.
- W4385572058 hasConceptScore W4385572058C101468663 @default.
- W4385572058 hasConceptScore W4385572058C127413603 @default.
- W4385572058 hasConceptScore W4385572058C137293760 @default.
- W4385572058 hasConceptScore W4385572058C154945302 @default.
- W4385572058 hasConceptScore W4385572058C173608175 @default.
- W4385572058 hasConceptScore W4385572058C177284502 @default.
- W4385572058 hasConceptScore W4385572058C199360897 @default.
- W4385572058 hasConceptScore W4385572058C201995342 @default.
- W4385572058 hasConceptScore W4385572058C204321447 @default.
- W4385572058 hasConceptScore W4385572058C2776175482 @default.
- W4385572058 hasConceptScore W4385572058C2776214188 @default.
- W4385572058 hasConceptScore W4385572058C2780451532 @default.
- W4385572058 hasConceptScore W4385572058C41008148 @default.
- W4385572058 hasConceptScore W4385572058C9390403 @default.
- W4385572058 hasLocation W43855720581 @default.
- W4385572058 hasOpenAccess W4385572058 @default.
- W4385572058 hasPrimaryLocation W43855720581 @default.
- W4385572058 hasRelatedWork W2112962394 @default.
- W4385572058 hasRelatedWork W2118300983 @default.
- W4385572058 hasRelatedWork W2359001871 @default.
- W4385572058 hasRelatedWork W2949992439 @default.
- W4385572058 hasRelatedWork W4205820553 @default.
- W4385572058 hasRelatedWork W4221160826 @default.
- W4385572058 hasRelatedWork W4243252198 @default.
- W4385572058 hasRelatedWork W4320486724 @default.
- W4385572058 hasRelatedWork W4378499352 @default.
- W4385572058 hasRelatedWork W4385567149 @default.
- W4385572058 isParatext "false" @default.
- W4385572058 isRetracted "false" @default.
- W4385572058 workType "article" @default.