Lemmatization criteria for the construction of Brazilian Sign Language lexicographic products

• Grant number: 21/02365-2
• Support Opportunities: Scholarships in Brazil - Post-Doctoral
• Start date: September 01, 2021
• End date: May 15, 2023
• Field of knowledge: Interdisciplinary Subjects
• Principal Investigator: José Mario de Martino
• Grantee: Antonielle Cantarelli Martins
• Host Institution: Faculdade de Engenharia Elétrica e de Computação (FEEC). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil

Abstract

This project aims to build a theoretical study blueprint to guide lemmatization processes in the phonetic-phonological, morphological, and semantic levels within Brazilian Sign Language - Libras. Lemmatization is understood as the grouping of lexical variants of the same stem within a Libras linguistic corpus in this project. Fourteen hundred thousand and five hundred (14.500) entries from Dic-Brasil: Dicionário da Língua de Sinais do Brasil: A Libras em suas mãos (Capovilla, Raphael, Temoteo, & Martins, 2017a, 2017b, 2017c) will be analyzed and Libras specific lemmatization principles will be designed to create lemmatized sign entry samples. The selected signs and their usage examples designed in this study will be performed by a native signer and registered via movement capture technology - MOCAP, according to the methodology described in De Martino et al. (2017). The corpus will be designed to be read by machines. The corpus notation system, that is, the linguistic descriptions, identification gloss systems, transcriptions, and translations, will be done using the software ELAN -EUDICO Linguistic Annotator. The expected outcomes of this project are: (1) a LIBRAS lemmatization theoretical model, (2) a Brazilian Portuguese/Libras parallel corpus, and (3) an online lemmatized bidirectional bilingual dictionary prototype. The efficacy of the lemmatization principles established in the study will be assessed through tests with a target group of dictionary users performing research exercises with the prototype. The use of lemmatization principles is expected to influence quantitative and qualitative descriptions of Libras lexical core in the future. Libras corpus with lemmatized entries contribute to Bilingual Deaf Education through the dissemination of Libras, Deaf people´s sign language and oral language better access, the promotion of linguistic analyses and tools for the teaching/learning practices, both in Libras for Deaf and hearing people, and Brazilian Portuguese language to Deaf people. (AU)

Scientific publications

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

DE MARTINO, JOSE MARIO; SILVA, IVANI RODRIGUES; MARQUES, JANICE GONCALVES TEMOTEO; MARTINS, ANTONIELLE CANTARELLI; POETA, ENZO TELLES; CHRISTINELE, DENER STASSUN; CAMPOS, JOAO PEDRO ARAUJO FERREIRA. Neural machine translation from text to sign language. UNIVERSAL ACCESS IN THE INFORMATION SOCIETY, v. N/A, p. 14-pg., 2023-07-18. (21/02365-2)