Low-resource neural machine translation with morphological modeling

arXiv:2404.02392v1 Announce Type: new
Abstract: Morphological modeling in neural machine translation (NMT) is a promising approach to achieving open-vocabulary machine translation for morphologically-rich languages. However, existing methods such as sub-word tokenization and character-based models are limited to the surface forms of the words. In this work, we propose a framework-solution for modeling complex morphology in low-resource settings. A two-tier transformer architecture is chosen to encode morphological information at the inputs. At the target-side output, a multi-task multi-label training scheme coupled …

abstract arxiv cs.cl forms framework however languages low machine machine translation modeling neural machine translation solution surface tokenization translation type word words work