Understanding and Addressing the Under-Translation Problem from the Perspective of Decoding Objective

TL;DR

This paper analyzes the root cause of under-translation in neural machine translation from the decoding objective perspective, proposing a confidence-based detection and correction method that improves translation completeness.

Contribution

It introduces a novel approach using EOS confidence as an under-translation detector and enhances penalties to address the issue effectively.

Findings

Accurately detects under-translation using EOS confidence

Improves translation completeness with minimal impact on correct outputs

Effective on both synthetic and real-world datasets

Abstract

Neural Machine Translation (NMT) has made remarkable progress over the past years. However, under-translation and over-translation remain two challenging problems in state-of-the-art NMT systems. In this work, we conduct an in-depth analysis on the underlying cause of under-translation in NMT, providing an explanation from the perspective of decoding objective. To optimize the beam search objective, the model tends to overlook words it is less confident about, leading to the under-translation phenomenon. Correspondingly, the model's confidence in predicting the End Of Sentence (EOS) diminishes when under-translation occurs, serving as a mild penalty for under-translated candidates. Building upon this analysis, we propose employing the confidence of predicting EOS as a detector for under-translation, and strengthening the confidence-based penalty to penalize candidates with a high risk of under-translation. Experiments on both synthetic and real-world data show that our method can accurately detect and rectify under-translated outputs, with minor impact on other correct translations.