Low-Resource Corpus Filtering using Multilingual Sentence Embeddings
This addresses the challenge of improving machine translation quality for low-resource languages like Nepali and Sinhala, though it is incremental as it builds on existing LASER methods with ensemble techniques.
The paper tackled the problem of filtering noisy parallel sentences for low-resource language pairs by using multilingual sentence embeddings from LASER without additional training, achieving the best performance in WMT19 tasks with BLEU score improvements of 1.3 and 1.4 over second-best systems.
In this paper, we describe our submission to the WMT19 low-resource parallel corpus filtering shared task. Our main approach is based on the LASER toolkit (Language-Agnostic SEntence Representations), which uses an encoder-decoder architecture trained on a parallel corpus to obtain multilingual sentence representations. We then use the representations directly to score and filter the noisy parallel sentences without additionally training a scoring function. We contrast our approach to other promising methods and show that LASER yields strong results. Finally, we produce an ensemble of different scoring methods and obtain additional gains. Our submission achieved the best overall performance for both the Nepali-English and Sinhala-English 1M tasks by a margin of 1.3 and 1.4 BLEU respectively, as compared to the second best systems. Moreover, our experiments show that this technique is promising for low and even no-resource scenarios.