Automatic Speech Recognition

  • Published 2017

Abstract

ASR is the first stage in an overall human/computer interaction pipeline that also includes Voicebox’s related Natural Language Understanding (NLU) and Text-to-Speech (TTS) technologies. Voicebox’s advanced ASR module is a multi-stage pipeline that uses techniques from machine learning, graph theory, traditional grammar development, and statistical analysis of large corpuses to form high-confidence transcriptions of input audio.

Cite this paper

@inproceedings{2017AutomaticSR, title={Automatic Speech Recognition}, author={}, year={2017} }