Semantic Lexical Grounding for Automatic Speech Recognition of Low-Resource West African Languages: Initial Experiments on Mina and Goun
Domain:
natural language processingRecord type:
paper
Creator:
djo
Publisher:
Host:
This record contains the original English manuscript and an official French translation of the paper "Semantic Lexical Grounding for Automatic Speech Recognition of Low-Resource West African Languages: Initial Experiments on Mina and Goun".
The paper presents the first documented automatic speech recognition (ASR) experiments for the Mina (Gen) and Goun (Gun/Gungbe) languages using Meta's MMS model. It introduces the Semantic Lexical Grounding (SLG) framework, a multilingual lexical resource of 4,708 verified concept pairs, and reports baseline ASR results obtained from community-collected speech corpora.
Files included:
- GOD_AI_arXiv_Paper_v2.pdf (English)
- GOD_AI_arXiv_Paper_v2_FR.pdf (Official French translation)
Similar
Semantic Lexical Grounding for Automatic Speech Recognition of Low-Resource West African Languages: Initial Experiments on Mina and Goun Ancrage Lexical Sémantique pour la Reconnaissance Automatique de la Parole des Langues d'Afrique de l'Ouest à Faibles Ressources : Expériences Initiales sur le Mina et le GounEnhancing Automatic Speech Recognition for Child Speech in Low-Resource LanguagesAutomatic Speech Recognition (ASR) for African Low-Resource Languages: A Systematic Literature Review