coded-by-49/langbridge_ai_v3_igbo_datasets

English ↔ Igbo Parallel Translation Dataset

Dataset Summary

A curated parallel corpus for English–Igbo machine translation, assembled from three high-quality, peer-reviewed sources. The dataset supports both translation directions (English → Igbo and Igbo → English) and was built specifically for fine-tuning NLLB-200 on this low-resource language pair.

Igbo is a Niger-Congo language spoken by over 50 million people, primarily in southeastern Nigeria, yet remains significantly underrepresented in NLP research and multilingual models.

---

Dataset Details

Split	Sentence Pairs
Train	12,009
Validation	1,693
Test	597
Total	14,299

Format: CSV with two columns — english and igbo.

---

Data Sources

This dataset aggregates Igbo–English parallel data from three reputable sources:

1. FLORES-101 — Goyal et al. (2022) Professional translations of sentences drawn from English Wikipedia, covering diverse general-knowledge topics. A standard benchmark for low-resource and multilingual machine translation evaluation.

2. Igbo-English Machine Translation Benchmark — Ezeani et al. (2020) A human-validated parallel corpus targeting the news domain, built specifically for Igbo as an evaluation benchmark. Constructed to address limitations in prior Igbo data sources such as religious-domain skew.

3. MAFAND-MT (African News Translation) — Adelani et al. (2022) A parallel news translation corpus for 16 African languages, created by fine-tuning pre-trained models with small quantities of high-quality translations. Igbo is among the supported languages.

---

Preprocessing

• Duplicates removed across all source datasets
• Data normalized into a unified CSV format
• Sources selected for human-translation quality and research credibility

---

Intended Use

This dataset is intended for:

• Fine-tuning multilingual translation models (e.g., NLLB-200) on the English–Igbo language pair
• Benchmarking machine translation performance for Igbo
• Low-resource NLP research on African languages

---

Limitations

All three source datasets are drawn from the news and encyclopedic domains. As a result:

• Informal, conversational, and colloquial Igbo is not represented
• Performance of models trained on this data may degrade on everyday speech, social media text, or dialectal variation
• Igbo tonal and diacritic complexity may be inconsistently represented across sources

---

Citation

If you use this dataset, please cite the original sources:

@article{goyal2022flores,
  title={The Flores-101 evaluation benchmark for low-resource and multilingual machine translation},
  author={Goyal, Naman and Gao, Cynthia and Chaudhary, Vishrav and Chen, Peng-Jen and Wenzek, Guillaume and Ju, Da and Krishnan, Sanjana and Ranzato, Marc'Aurelio and Guzm{\'a}n, Francisco and Fan, Angela},
  journal={Transactions of the Association for Computational Linguistics},
  volume={10},
  pages={522--538},
  year={2022},
  publisher={MIT Press}
}

@misc{ezeani2020igboenglish,
  title={Igbo-English Machine Translation: An Evaluation Benchmark},
  author={Ignatius Ezeani and Paul Rayson and Ikechukwu Onyenwe and Chinedu Uchechukwu and Mark Hepple},
  year={2020},
  eprint={2004.00648},
  archivePrefix={arXiv},
  primaryClass={cs.CL},
  url={https://arxiv.org/abs/2004.00648}
}

@inproceedings{adelani-etal-2022-thousand,
  title={A Few Thousand Translations Go a Long Way! Leveraging Pre-trained Models for {A}frican News Translation},
  author={Adelani, David and Alabi, Jesujoba and Fan, Angela and Kreutzer, Julia and others},
  booktitle={Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies},
  pages={3053--3070},
  year={2022},
  publisher={Association for Computational Linguistics},
  url={https://aclanthology.org/2022.naacl-main.223}
}

@article{nllb2022,
  author={NLLB Team, Marta R. Costa-jussà, James Cross, Onur Çelebi, Maha Elbayad, Kenneth Heafield, Kevin Heffernan, Elahe Kalbassi, Janice Lam, Daniel Licht, Jean Maillard, Anna Sun, Skyler Wang, Guillaume Wenzek, Al Youngblood, Bapi Akula, Loic Barrault, Gabriel Mejia Gonzalez, Prangthip Hansanti, John Hoffman, Semarley Jarrett, Kaushik Ram Sadagopan, Dirk Rowe, Shannon Spruit, Chau Tran, Pierre Andrews, Necip Fazil Ayan, Shruti Bhosale, Sergey Edunov, Angela Fan, Cynthia Gao, Vedanuj Goswami, Francisco Guzmán, Philipp Koehn, Alexandre Mourachko, Christophe Ropers, Safiyyah Saleem, Holger Schwenk, Jeff Wang},
  title={No Language Left Behind: Scaling Human-Centered Machine Translation},
  year={2022}
}