Abstract
In the 1960s, automatic speech recognition has been widely studied. In the past, HMM has been the mainstream of the acoustic model. With the development of machine learning, neural network is introduced to the speech recognition, relying on neural network’s strong learning ability. Thus the acoustic model of DNN has significantly improved the voice recognition rate, compared to the HMM model. In order to simplify the traditional speech recognition system, the end-to-end speech recognition method is proposed. This paper mainly introduces and analyzes the end-to-end system, and the main two models of CTC and attention, as well as the prospect of future speech recognition research.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
