Deep Learning Based Natural Language Processing for End to End Speech Translation
share
Deep Learning methods employ multiple processing layers to learn hierarchial representations of data. They have already been deployed in a humongous number of applications and have produced state-of-the-art results. Recently with the growth in processing power of computers to be able to do high dimensional tensor calculations, Natural Language Processing (NLP) applications have been given a significant boost in terms of efficiency as well as accuracy. In this paper, we will take a look at various signal processing techniques and then application of them to produce a speech-to-text system using Deep Recurrent Neural Networks.
READ FULL TEXT