NICT's Corpus Filtering Systems for the WMT18 Parallel Corpus Filtering Task

TL;DR

This paper describes NICT's approach to filtering noisy web-crawled data for neural machine translation by designing features and training classifiers, resulting in improved translation systems.

Contribution

NICT developed a feature-based classifier to filter noisy web data, enabling the creation of cleaner datasets for better NMT performance.

Findings

Filtered data improved NMT translation quality

Sampling 100M and 10M words yielded promising results

Classifier effectively identified high-quality sentence pairs

Abstract

This paper presents the NICT's participation in the WMT18 shared parallel corpus filtering task. The organizers provided 1 billion words German-English corpus crawled from the web as part of the Paracrawl project. This corpus is too noisy to build an acceptable neural machine translation (NMT) system. Using the clean data of the WMT18 shared news translation task, we designed several features and trained a classifier to score each sentence pairs in the noisy data. Finally, we sampled 100 million and 10 million words and built corresponding NMT systems. Empirical results show that our NMT systems trained on sampled data achieve promising performance.