Logo Lanfrica
  • Accueil
  • Atlas
  • Analyses
  • Documentation
  • Sign in

© 2026 Lanfrica. Tous droits réservés. Tous les droits d'auteur des ressources affichées sur le site Web Lanfrica appartiennent aux détenteurs de droits d'auteur d'origine, sauf indication contraire explicite.

Mafoko: Structuring and Building Open Multilingual Terminologies for South African NLP

Domaine:

natural language processing

Type de record:

datasetpaper
Créateur:
Vukosi Marivate

The critical lack of structured terminological data for South Africa's official languages hampers progress in multilingual NLP, despite the existence of numerous government and academic terminology lists. These valuable assets remain fragmented and locked in non-machine-readable formats, rendering them unusable for computational research and development.

Mafoko addresses this challenge by systematically aggregating, cleaning, and standardising these scattered resources into open, interoperable datasets.

We introduce the foundational Mafoko dataset, released under the equitable, Africa-centered NOODL framework. To demonstrate its immediate utility, we integrate the terminology into a Retrieval-Augmented Generation (RAG) pipeline. Experiments show substantial improvements in the accuracy and domain-specific consistency of English-to-Tshivenda machine translation for large language models. Mafoko provides a scalable foundation for developing robust and equitable NLP technologies, ensuring South Africa's rich linguistic diversity is represented in the digital age.

Presented at the Sixth Workshop on Resources for African Indigenous Languages (RAIL 2025).

Visit

arxiv.orgproject pagehuggingface.co

Tasks

machine translation

Languages

AfrikaansBirwaNdebeleNdebeleSetswanaSotho, NorthernSwatiTsongaVendaXhosa+1

Tags

terminology datasetlinguistic resourcelinguisticsmultilingual NLP datasetterminology managementmachine translationlanguage resource standardisationAfrican language technologiesSouth AfricaSouth African languages+2

Licenses

https://licensingafricandatasets.com/nwulite-obodo-licenseCC-BY-NC-SA 2.5 ZA

Similaires

BUILDING ROBUST NLP SYSTEMS FOR AFRICAN LANGUAGES USING TRANSFER LEARNING AND MULTILINGUAL EMBEDDINGSdsfsi/za-mafoko-unisa-multilingualBuilding OMProDat: an open multilingual prosodic databaseMultilingual NLP for African Healthcare: Bias, Translation, and Explainability ChallengesUniplexity-AI/-Building-Multilingual-AI-Systems-for-African-BusinessesSouth African Open Access and Open Data

BUILDING ROBUST NLP SYSTEMS FOR AFRICAN LANGUAGES USING TRANSFER LEARNING AND MULTILINGUAL EMBEDDINGS

dsfsi/za-mafoko-unisa-multilingual

The South African Multilingual Linguistic Terminology (SAMLT) Project is a comprehensive multilingua

Building OMProDat: an open multilingual prosodic database

International audience Current research on speech prosody generally makes use of larg

Multilingual NLP for African Healthcare: Bias, Translation, and Explainability Challenges

Uniplexity-AI/-Building-Multilingual-AI-Systems-for-African-Businesses

Building Multilingual AI Systems for African Businesses: GraphRAG and GNN-Based Fine-Tuning for Engl

South African Open Access and Open Data

Join our keynotes Daisy Salematsela (Director, WITS Libraries, Johannesburg, South Africa) & Laz