AfroMT: Pretraining Strategies and Reproducible Benchmarks for Translation of 8 African Languages
This addresses the problem of limited translation resources for millions of speakers of African languages, though it is incremental in extending existing methods to new data.
The paper tackles the lack of standardized benchmarks for African languages in machine translation by introducing AfroMT, a reproducible benchmark for eight languages, and shows that novel pretraining strategies yield gains of up to 2 BLEU points over baselines and up to 12 BLEU points in data-constrained scenarios.
Reproducible benchmarks are crucial in driving progress of machine translation research. However, existing machine translation benchmarks have been mostly limited to high-resource or well-represented languages. Despite an increasing interest in low-resource machine translation, there are no standardized reproducible benchmarks for many African languages, many of which are used by millions of speakers but have less digitized textual data. To tackle these challenges, we propose AfroMT, a standardized, clean, and reproducible machine translation benchmark for eight widely spoken African languages. We also develop a suite of analysis tools for system diagnosis taking into account the unique properties of these languages. Furthermore, we explore the newly considered case of low-resource focused pretraining and develop two novel data augmentation-based strategies, leveraging word-level alignment information and pseudo-monolingual data for pretraining multilingual sequence-to-sequence models. We demonstrate significant improvements when pretraining on 11 languages, with gains of up to 2 BLEU points over strong baselines. We also show gains of up to 12 BLEU points over cross-lingual transfer baselines in data-constrained scenarios. All code and pretrained models will be released as further steps towards larger reproducible benchmarks for African languages.