Improving Hybrid CTC/Attention Architecture with Time-Restricted Self-Attention CTC for End-to-End Speech Recognition

Improving Hybrid CTC/Attention Architecture with Time-Restricted Self-Attention CTC for End-to-End Speech Recognition

As demonstrated in hybrid connectionist temporal classification (CTC)/Attention architecture, joint training with a CTC objective is very effective to solve the misalignment problem existing in the attention-based end-to-end automatic speech recognition (ASR) framework. However, the CTC output relie...

Full description

Bibliographic Details
Main Authors: Long Wu, Ta Li, Li Wang, Yonghong Yan
Format: Article
Language:English
Published: MDPI AG 2019-10-01
Series:Applied Sciences
Subjects:
automatic speech recognition
end-to-end
ctc
self-attention
hybrid ctc/attention
Online Access:https://www.mdpi.com/2076-3417/9/21/4639

Internet

https://www.mdpi.com/2076-3417/9/21/4639

Similar Items