Exploring the planet of the APEs: A comparative study of state-of-the-art methods for MT automatic post-editing

Abstract

Downstream processing of machine trans-lation (MT) output promises to be a so-lution to improve translation quality, es-pecially when the MT system's internal decoding process is not accessible. Both rule-based and statistical automatic post-editing (APE) methods have been pro-posed over the years, but with contrast-ing results. A missing aspect in previous evaluations is the assessment of different methods: i) under comparable conditions, and ii) on different language pairs featur-ing variable levels of MT quality. Fo-cusing on statistical APE methods (more portable across languages), we propose the first systematic analysis of two ap-proaches. To understand their potential, we compare them in the same conditions over six language pairs having English as source. Our results evidence consis-tent improvements on all language pairs, a relation between the extent of the gain and MT output quality, slight but statis-tically significant performance differences between the two methods, and their possi-ble complementarity.