Extracting Parallel Sentences with Bidirectional Recurrent Neural Networks to Improve Machine Translation

TL;DR

This paper introduces a bidirectional RNN method for extracting parallel sentences from multilingual texts, improving machine translation without relying on feature engineering or external resources.

Contribution

The paper presents a novel bidirectional RNN approach for parallel sentence extraction that outperforms baselines and enhances translation quality.

Findings

Achieved promising extraction results from noisy corpora

Improved machine translation performance using extracted sentence pairs

Eliminated need for feature engineering or external resources

Abstract

Parallel sentence extraction is a task addressing the data sparsity problem found in multilingual natural language processing applications. We propose a bidirectional recurrent neural network based approach to extract parallel sentences from collections of multilingual texts. Our experiments with noisy parallel corpora show that we can achieve promising results against a competitive baseline by removing the need of specific feature engineering or additional external resources. To justify the utility of our approach, we extract sentence pairs from Wikipedia articles to train machine translation systems and show significant improvements in translation performance.