Can Automatic Post-Editing Improve NMT?

TL;DR

This paper demonstrates that training neural automatic post-editing models on a large, human-annotated corpus can significantly improve neural machine translation outputs, challenging previous assumptions about APE's limited effectiveness.

Contribution

The authors compile a large human post-editing corpus for English-German NMT and show that neural APE models trained on this data can substantially enhance translation quality.

Findings

Neural APE models trained on human data improve NMT outputs.

Artificial training data and domain adaptation impact APE performance.

Large, high-quality corpora are crucial for effective neural APE.

Abstract

Automatic post-editing (APE) aims to improve machine translations, thereby reducing human post-editing effort. APE has had notable success when used with statistical machine translation (SMT) systems but has not been as successful over neural machine translation (NMT) systems. This has raised questions on the relevance of APE task in the current scenario. However, the training of APE models has been heavily reliant on large-scale artificial corpora combined with only limited human post-edited data. We hypothesize that APE models have been underperforming in improving NMT translations due to the lack of adequate supervision. To ascertain our hypothesis, we compile a larger corpus of human post-edits of English to German NMT. We empirically show that a state-of-art neural APE model trained on this corpus can significantly improve a strong in-domain NMT system, challenging the current understanding in the field. We further investigate the effects of varying training data sizes, using artificial training data, and domain specificity for the APE task. We release this new corpus under CC BY-NC-SA 4.0 license at https://github.com/shamilcm/pedra.