Attaining the Unattainable? Reassessing Claims of Human Parity in Neural Machine Translation

TLDR

The authors reexamine a 2018 claim of human parity in Chinese‑to‑English news translation by incorporating source language, evaluator proficiency, and inter‑sentential context, and they propose guidelines for future human evaluations. They restrict analysis to original source text, compare professional translators with non‑experts, and examine human translations for key issues using pairwise ranking. Their results show that human parity is not achieved on original source text, that expert judgments yield higher agreement and better discrimination, and they recommend improved evaluation practices.

Abstract

We reassess a recent study (Hassan et al., 2018) that claimed that machine translation (MT) has reached human parity for the translation of news from Chinese into English, using pairwise ranking and considering three variables that were not taken into account in that previous study: the language in which the source side of the test set was originally written, the translation proficiency of the evaluators, and the provision of inter-sentential context. If we consider only original source text (i.e. not translated from another language, or translationese), then we find evidence showing that human parity has not been achieved. We compare the judgments of professional translators against those of non-experts and discover that those of the experts result in higher inter-annotator agreement and better discrimination between human and machine translations. In addition, we analyse the human translations of the test set and identify important translation issues. Finally, based on these findings, we provide a set of recommendations for future human evaluations of MT.