Did Translation Models Get More Robust Without Anyone Even Noticing?

arXiv:2403.03923v1 Announce Type: new
Abstract: Neural machine translation (MT) models achieve strong results across a variety of settings, but it is widely believed that they are highly sensitive to "noisy" inputs, such as spelling errors, abbreviations, and other formatting issues. In this paper, we revisit this insight in light of recent multilingual MT models and large language models (LLMs) applied to machine translation. Somewhat surprisingly, we show through controlled experiments that these models are far more robust to many kinds …

abstract arxiv cs.cl errors inputs insight light machine machine translation neural machine translation paper results robust translation type