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

2025-08-05

AI Summary

South Africa’s official languages have long lacked structured, machine-readable terminology resources, severely hindering multilingual natural language processing (NLP) development. Method: This paper introduces the first African-centered, open-source multilingual terminology framework—built on NOODL—that systematically integrates, cleans, and standardizes fragmented, unstructured term lists from government and academic institutions into an open, interoperable multilingual terminology repository. It innovatively incorporates retrieval-augmented generation (RAG) to dynamically inject structured terminology into large language model (LLM)-based translation pipelines. Contribution/Results: Experiments demonstrate significant improvements in accuracy and domain consistency for English-to-Tsonga translation, validating the terminology repository’s critical role in advancing equitable, locally grounded NLP technologies. The framework bridges infrastructural gaps in African language NLP and supports scalable, context-aware, and terminologically precise machine translation.

Technology Category

• Natural Language Processing: Machine Translation, Multilinguality, Cross-Lingual NLP
• Machine Learning: Large Multimodal Models (LMMs)
• Knowledge Representation and Reasoning: Knowledge Representation Languages

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data

Abstract

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. emph{Marito} addresses this challenge by systematically aggregating, cleaning, and standardising these scattered resources into open, interoperable datasets. We introduce the foundational emph{Marito} 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. emph{Marito} provides a scalable foundation for developing robust and equitable NLP technologies, ensuring South Africa's rich linguistic diversity is represented in the digital age.

Problem

• Lack of structured multilingual terminologies for South African NLP
• Fragmented terminological data in non-machine-readable formats
• Need for scalable foundation for equitable NLP technologies

Innovation

• Aggregating and standardizing multilingual terminological data
• Releasing datasets under Africa-centered NOODL framework
• Integrating terminology into Retrieval-Augmented Generation pipeline