Achieving Human Parity on Automatic Chinese to English News Translation

TL;DR

This paper demonstrates that Microsoft's neural machine translation system for Chinese to English news translation has achieved human parity in translation quality, surpassing non-professional human translations.

Contribution

It introduces a method to measure human parity and shows that a neural machine translation system can reach this level on a major benchmark.

Findings

Neural machine translation system reaches human parity in quality.

System significantly outperforms crowd-sourced non-professional translations.

Achieves state-of-the-art results on WMT 2017 Chinese-English news translation.

Abstract

Machine translation has made rapid advances in recent years. Millions of people are using it today in online translation systems and mobile applications in order to communicate across language barriers. The question naturally arises whether such systems can approach or achieve parity with human translations. In this paper, we first address the problem of how to define and accurately measure human parity in translation. We then describe Microsoft's machine translation system and measure the quality of its translations on the widely used WMT 2017 news translation task from Chinese to English. We find that our latest neural machine translation system has reached a new state-of-the-art, and that the translation quality is at human parity when compared to professional human translations. We also find that it significantly exceeds the quality of crowd-sourced non-professional translations.