End-to-End Speech Recognition

Deep Speech

While at Baidu Research, I had the privilege of working on the revolutionary Deep Speech end-to-end speech recognition system. Deep Speech did away with the complicated traditional speech recognition pipeline, replacing it instead with a large neural network that is trained in an end-to-end fashion to convert audio into text.

Publications

. Exploring Neural Transducers for End-to-End Speech Recognition. ASRU, 2017.

arXiv PDF Project Ref

. Reducing Bias in Production Speech Models. arXiv, 2017.

arXiv PDF Project

. Deep Speech 2: End-to-End Speech Recognition in English and Mandarin. ICML, 2016.

arXiv PDF Project Slides Ref