Exploiting Language Relatedness in Machine Translation Through Domain Adaptation Techniques
This work addresses the problem of data scarcity in machine translation for low-resource languages, offering incremental improvements through domain adaptation.
The paper tackled the challenge of improving machine translation quality for low-resource languages by proposing a novel approach using a scaled similarity score based on a 5-gram KenLM language model to filter in-domain data, combined with domain adaptation techniques. The result was an increase of approximately 2-3 BLEU points across different methods for the Hindi-Nepali language pair.
One of the significant challenges of Machine Translation (MT) is the scarcity of large amounts of data, mainly parallel sentence aligned corpora. If the evaluation is as rigorous as resource-rich languages, both Neural Machine Translation (NMT) and Statistical Machine Translation (SMT) can produce good results with such large amounts of data. However, it is challenging to improve the quality of MT output for low resource languages, especially in NMT and SMT. In order to tackle the challenges faced by MT, we present a novel approach of using a scaled similarity score of sentences, especially for related languages based on a 5-gram KenLM language model with Kneser-ney smoothing technique for filtering in-domain data from out-of-domain corpora that boost the translation quality of MT. Furthermore, we employ other domain adaptation techniques such as multi-domain, fine-tuning and iterative back-translation approach to compare our novel approach on the Hindi-Nepali language pair for NMT and SMT. Our approach succeeds in increasing ~2 BLEU point on multi-domain approach, ~3 BLEU point on fine-tuning for NMT and ~2 BLEU point on iterative back-translation approach.