Igbo-English Machine Translation: An Evaluation Benchmark

TL;DR

The paper tackles the lack of Igbo NLP resources by building a standard Igbo–English MT benchmark. It details a four-phase workflow for data collection, human-in-the-loop translation, quality control, and monolingual data assembly, resulting in 11,584 parallel sentence pairs and ~383k monolingual Igbo sentences. Data are openly available under Creative Commons on GitHub to enable reproducible, fair model comparisons and broader community use. By critiquing existing sources like OPUS and JW.ORG for evaluation, the work lays groundwork for robust assessment of Igbo MT and informs future MT developments for a low-resource language.

Abstract

Although researchers and practitioners are pushing the boundaries and enhancing the capacities of NLP tools and methods, works on African languages are lagging. A lot of focus on well resourced languages such as English, Japanese, German, French, Russian, Mandarin Chinese etc. Over 97% of the world's 7000 languages, including African languages, are low resourced for NLP i.e. they have little or no data, tools, and techniques for NLP research. For instance, only 5 out of 2965, 0.19% authors of full text papers in the ACL Anthology extracted from the 5 major conferences in 2018 ACL, NAACL, EMNLP, COLING and CoNLL, are affiliated to African institutions. In this work, we discuss our effort toward building a standard machine translation benchmark dataset for Igbo, one of the 3 major Nigerian languages. Igbo is spoken by more than 50 million people globally with over 50% of the speakers are in southeastern Nigeria. Igbo is low resourced although there have been some efforts toward developing IgboNLP such as part of speech tagging and diacritic restoration