: Vocabulario de PLN: low-resource language

computational linguistics > sociolinguistics > multilingualism > low-resource language

responsible NLP > low-resource language

low-resource language

A language for which there are few, if any, documenting resources such as lexicons, grammars, or written texts. (Lane, Natural Language Processing for Low Resource Languages (project description), Charles Darwin University)

Low-resource languages are underrepresented in pre-trained multilingual LMs. (Hangya, Saadi & Fraser, 2022)
Our approach does not assume any knowledge about foreign languages making it applicable to a wide range of resource-poor languages. (Zennaki, Semmar & Besacier, 2015)
Secondly there is no transfer learning between language pairs due to separate models which means that low-resource languages generally have low translation quality. (Purason & Tättar, 2022)
Then we randomly selected 29000 lines form the bitext and constructed a reduced training corpus to simulate a resource-poor language. (He, Zhao, Wang & Liu, 2011)
To try to strengthen the baselines for underrepresented languages we perform vocabulary extension and model fine-tuning using the monolingual data of the target low-resource language. (Hangya, Saadi & Fraser, 2022)

http://data.loterre.fr/ark:/67375/8LP-BG78HVF2-9

RDF/XML TURTLE JSON-LD última modificación 26/6/24

Loterre