Feriji: A French-Zarma Parallel Corpus, Glossary & Translator

This dataset repository contains Feriji, the first robust French-Zarma parallel corpus and glossary designed for machine translation (MT), as introduced in the paper "Feriji: A French-Zarma Parallel Corpus, Glossary & Translator" published at ACL 2024 (Student Research Workshop). Zarma, a dialect of Songhay within the Nilo-Saharan language family, is spoken by over 5 million people across Niger and neighboring countries. Feriji addresses the critical need for linguistic resources for this underrepresented African language.

Dataset Overview

The Feriji dataset includes three components:

• French-Zarma Parallel Corpus: Located in corpus/fr_dje.csv, this contains 61,085 sentences in Zarma and 42,789 in French, forming a robust aligned corpus for MT.
• French-Zarma Glossary: Found in glossary/glossary.csv, this includes 4,062 word pairs, serving as a lexical resource.
• Monolingual Zarma Text: Stored in monolingual/monolingual_zarma.txt, this provides additional Zarma-only text for language modeling or analysis.

These resources were used to fine-tune three large language models, achieving a BLEU score of 30.06 on the best-performing model, with further evaluation based on human judgments of fluency, comprehension, and readability.

Repository Structure

• corpus/fr_dje.csv: Aligned French-Zarma sentence pairs.
• glossary/glossary.csv: French-Zarma word glossary.
• monolingual/monolingual_zarma.txt: Monolingual Zarma text.

Paper Details

For full details on the creation, evaluation, and impact of Feriji, see the original paper:

• Link: https://aclanthology.org/2024.acl-srw.1/
• Citation:

  @inproceedings{keita-etal-2024-feriji,
      title = "Feriji: A {F}rench-{Z}arma Parallel Corpus, Glossary {\&} Translator",
      author = "Keita, Mamadou  and
        Ibrahim, Elysabhete  and
        Alfari, Habibatou  and
        Homan, Christopher",
      editor = "Fu, Xiyan  and
        Fleisig, Eve",
      booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 4: Student Research Workshop)",
      month = aug,
      year = "2024",
      address = "Bangkok, Thailand",
      publisher = "Association for Computational Linguistics",
      url = "https://aclanthology.org/2024.acl-srw.1/",
      doi = "10.18653/v1/2024.acl-srw.1",
      pages = "1--9",
      abstract = "Machine translation (MT) is a rapidly expanding field that has experienced significant advancements in recent years with the development of models capable of translating multiple languages with remarkable accuracy. However, the representation of African languages in this field still needs improvement due to linguistic complexities and limited resources. This applies to the Zarma language, a dialect of Songhay (of the Nilo-Saharan language family) spoken by over 5 million people across Niger and neighboring countries (Lewis et al., 2016). This paper introduces Feriji, the first robust French-Zarma parallel corpus and glossary designed for MT. The corpus, containing 61,085 sentences in Zarma and 42,789 in French, and a glossary of 4,062 words represents a significant step in addressing the need for more resources for Zarma. We fine-tune three large language models on our dataset, obtaining a BLEU score of 30.06 on the best-performing model. We further evaluate the models on human judgments of fluency, comprehension, and readability and the importance and impact of the corpus and models. Our contributions help to bridge a significant language gap and promote an essential and overlooked indigenous African language."
  }