A portable English translation system through deep learning and internet of things

Abstract

Portable English translation systems are increasingly important in daily life. Existing online English translation based on Transformer has become mainstream on the service side. But Transformer loses the ability to capture local dependencies, leading to local syntax errors. The graph convolution operation can fully model the local relationship. Therefore, this paper proposes a new graph embedded Transformer network (GETN) for Portable English translation based on Internet of Things technology. Specifically, the portable device converts the voice signal into text, and then transmits it to the server through Internet technology. By introducing the graph convolution module into the existing Transformer, local relations can be effectively modeled. The results of the existing English grammatical error correction dataset show that the graph-embedded mechanism can achieve a higher translation effect with fewer parameters and effectively alleviate local syntax errors.