Matches in SemOpenAlex for { <https://semopenalex.org/work/W3168264548> ?p ?o ?g. }
- W3168264548 abstract "Machine translation (MT) is currently evaluated in one of two ways: in a monolingual fashion, by comparison with the system output to one or more human reference translations, or in a trained crosslingual fashion, by building a supervised model to predict quality scores from human-labeled data. In this paper, we propose a more cost-effective, yet well performing unsupervised alternative SentSim: relying on strong pretrained multilingual word and sentence representations, we directly compare the source with the machine translated sentence, thus avoiding the need for both reference translations and labelled training data. The metric builds on state-of-the-art embedding-based approaches – namely BERTScore and Word Mover’s Distance – by incorporating a notion of sentence semantic similarity. By doing so, it achieves better correlation with human scores on different datasets. We show that it outperforms these and other metrics in the standard monolingual setting (MT-reference translation), a well as in the source-MT bilingual setting, where it performs on par with glass-box approaches to quality estimation that rely on MT model information." @default.
- W3168264548 created "2021-06-22" @default.
- W3168264548 creator A5052281233 @default.
- W3168264548 creator A5053217291 @default.
- W3168264548 creator A5082442045 @default.
- W3168264548 date "2021-01-01" @default.
- W3168264548 modified "2023-10-17" @default.
- W3168264548 title "SentSim: Crosslingual Semantic Evaluation of Machine Translation" @default.
- W3168264548 cites W1614298861 @default.
- W3168264548 cites W1840435438 @default.
- W3168264548 cites W1956340063 @default.
- W3168264548 cites W1975742580 @default.
- W3168264548 cites W2101105183 @default.
- W3168264548 cites W2123301721 @default.
- W3168264548 cites W2143668817 @default.
- W3168264548 cites W2154652894 @default.
- W3168264548 cites W2164984707 @default.
- W3168264548 cites W2313299119 @default.
- W3168264548 cites W2564590796 @default.
- W3168264548 cites W2756675635 @default.
- W3168264548 cites W2760738985 @default.
- W3168264548 cites W2915756181 @default.
- W3168264548 cites W2945405384 @default.
- W3168264548 cites W2947771965 @default.
- W3168264548 cites W2952215760 @default.
- W3168264548 cites W2953287808 @default.
- W3168264548 cites W2963341956 @default.
- W3168264548 cites W2963846996 @default.
- W3168264548 cites W2970597249 @default.
- W3168264548 cites W2970785793 @default.
- W3168264548 cites W2970986500 @default.
- W3168264548 cites W2970986510 @default.
- W3168264548 cites W2971094522 @default.
- W3168264548 cites W2996403597 @default.
- W3168264548 cites W3035252911 @default.
- W3168264548 cites W3035390927 @default.
- W3168264548 cites W3084095723 @default.
- W3168264548 cites W3100806282 @default.
- W3168264548 cites W3103901889 @default.
- W3168264548 cites W3104033643 @default.
- W3168264548 cites W3119107795 @default.
- W3168264548 cites W3119878165 @default.
- W3168264548 cites W3119881489 @default.
- W3168264548 cites W3120791402 @default.
- W3168264548 cites W658020064 @default.
- W3168264548 doi "https://doi.org/10.18653/v1/2021.naacl-main.252" @default.
- W3168264548 hasPublicationYear "2021" @default.
- W3168264548 type Work @default.
- W3168264548 sameAs 3168264548 @default.
- W3168264548 citedByCount "11" @default.
- W3168264548 countsByYear W31682645482021 @default.
- W3168264548 countsByYear W31682645482022 @default.
- W3168264548 countsByYear W31682645482023 @default.
- W3168264548 crossrefType "proceedings-article" @default.
- W3168264548 hasAuthorship W3168264548A5052281233 @default.
- W3168264548 hasAuthorship W3168264548A5053217291 @default.
- W3168264548 hasAuthorship W3168264548A5082442045 @default.
- W3168264548 hasBestOaLocation W31682645481 @default.
- W3168264548 hasConcept C103278499 @default.
- W3168264548 hasConcept C104317684 @default.
- W3168264548 hasConcept C105580179 @default.
- W3168264548 hasConcept C111472728 @default.
- W3168264548 hasConcept C115961682 @default.
- W3168264548 hasConcept C130318100 @default.
- W3168264548 hasConcept C138885662 @default.
- W3168264548 hasConcept C149364088 @default.
- W3168264548 hasConcept C154945302 @default.
- W3168264548 hasConcept C162324750 @default.
- W3168264548 hasConcept C176217482 @default.
- W3168264548 hasConcept C185592680 @default.
- W3168264548 hasConcept C203005215 @default.
- W3168264548 hasConcept C204321447 @default.
- W3168264548 hasConcept C21547014 @default.
- W3168264548 hasConcept C2777462759 @default.
- W3168264548 hasConcept C2777530160 @default.
- W3168264548 hasConcept C2779530757 @default.
- W3168264548 hasConcept C41008148 @default.
- W3168264548 hasConcept C41608201 @default.
- W3168264548 hasConcept C41895202 @default.
- W3168264548 hasConcept C55493867 @default.
- W3168264548 hasConcept C90805587 @default.
- W3168264548 hasConceptScore W3168264548C103278499 @default.
- W3168264548 hasConceptScore W3168264548C104317684 @default.
- W3168264548 hasConceptScore W3168264548C105580179 @default.
- W3168264548 hasConceptScore W3168264548C111472728 @default.
- W3168264548 hasConceptScore W3168264548C115961682 @default.
- W3168264548 hasConceptScore W3168264548C130318100 @default.
- W3168264548 hasConceptScore W3168264548C138885662 @default.
- W3168264548 hasConceptScore W3168264548C149364088 @default.
- W3168264548 hasConceptScore W3168264548C154945302 @default.
- W3168264548 hasConceptScore W3168264548C162324750 @default.
- W3168264548 hasConceptScore W3168264548C176217482 @default.
- W3168264548 hasConceptScore W3168264548C185592680 @default.
- W3168264548 hasConceptScore W3168264548C203005215 @default.
- W3168264548 hasConceptScore W3168264548C204321447 @default.
- W3168264548 hasConceptScore W3168264548C21547014 @default.
- W3168264548 hasConceptScore W3168264548C2777462759 @default.
- W3168264548 hasConceptScore W3168264548C2777530160 @default.
- W3168264548 hasConceptScore W3168264548C2779530757 @default.
- W3168264548 hasConceptScore W3168264548C41008148 @default.