Edge NLP for Efficient Machine Translation in Low Connectivity Areas

Machine translation (MT) usually requires connectivity and access to the cloud which is often limited in many parts of the world, including hard to reach rural areas. Natural language processing (NLP) on the edge aims to solve this problem by processing language data closer to the source. To achieve this, 100 sentence pairs were stored and processed on a Raspberry Pi, and a recurrent neural network (RNN) using the long short-term memory (LSTM) architecture was used for MT. We are focusing on translating between English and Hausa, a low-resource language spoken in West Africa. It was found that the developed prototype produced “good and fluent translations” with a training accuracy of 91%. The model also achieved a BLEU score of 73.5, compared to the existing models that have scores of 22.2 and below.