Using LSTM to Translate French to Senegalese Local Languages: Wolof as a Case Study
This addresses translation for Wolof speakers, but it is incremental as it applies existing neural methods to a new low-resource language.
The paper tackled machine translation for Wolof, a low-resource language, by developing an LSTM-based system with attention, achieving a 47% BLEU score on a dataset of 35,000 parallel sentences.
In this paper, we propose a neural machine translation system for Wolof, a low-resource Niger-Congo language. First we gathered a parallel corpus of 70000 aligned French-Wolof sentences. Then we developped a baseline LSTM based encoder-decoder architecture which was further extended to bidirectional LSTMs with attention mechanisms. Our models are trained on a limited amount of parallel French-Wolof data of approximately 35000 parallel sentences. Experimental results on French-Wolof translation tasks show that our approach produces promising translations in extremely low-resource conditions. The best model was able to achieve a good performance of 47% BLEU score.