Low-Resource Corpus Filtering Using Multilingual Sentence Embeddings
Vishrav Chaudharya b, Yuqing Tanga b, Francisco Guzmána b, Holger Schwenka b, Philipp Koehna c
- aFacebook AI Johns Hopkins University
- bFacebook
- cJohns Hopkins University
Papers open inside SciRah from the original publisher or repository; SciRah never hosts them.
Abstract
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.
TL;DRThe 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 to score and filter the noisy parallel sentences without additionally training a scoring function. (Semantic Scholar)