Improving Neural Machine Translation of Indigenous Languages with Multilingual Transfer Learning
This work addresses the problem of low-resource machine translation for Indigenous languages, which is crucial for preserving endangered languages, though it is incremental in its approach.
The paper tackled the challenge of machine translation for Indigenous languages with limited parallel data by using multilingual transfer learning from Spanish to ten South American languages, achieving new state-of-the-art results on five language pairs, including doubling performance on one pair.
Machine translation (MT) involving Indigenous languages, including those possibly endangered, is challenging due to lack of sufficient parallel data. We describe an approach exploiting bilingual and multilingual pretrained MT models in a transfer learning setting to translate from Spanish to ten South American Indigenous languages. Our models set new SOTA on five out of the ten language pairs we consider, even doubling performance on one of these five pairs. Unlike previous SOTA that perform data augmentation to enlarge the train sets, we retain the low-resource setting to test the effectiveness of our models under such a constraint. In spite of the rarity of linguistic information available about the Indigenous languages, we offer a number of quantitative and qualitative analyses (e.g., as to morphology, tokenization, and orthography) to contextualize our results.