Attentive fine-tuning of Transformers for Translation of low-resourced languages @LoResMT 2021
This work addresses translation challenges for low-resourced languages, but it is incremental as it applies existing fine-tuning methods to new language pairs.
The paper tackled machine translation for low-resourced languages like Marathi and Irish by fine-tuning pretrained models, achieving top rankings in shared tasks with promising BLEU scores.
This paper reports the Machine Translation (MT) systems submitted by the IIITT team for the English->Marathi and English->Irish language pairs LoResMT 2021 shared task. The task focuses on getting exceptional translations for rather low-resourced languages like Irish and Marathi. We fine-tune IndicTrans, a pretrained multilingual NMT model for English->Marathi, using external parallel corpus as input for additional training. We have used a pretrained Helsinki-NLP Opus MT English->Irish model for the latter language pair. Our approaches yield relatively promising results on the BLEU metrics. Under the team name IIITT, our systems ranked 1, 1, and 2 in English->Marathi, Irish->English, and English->Irish, respectively.