Effective Exploitation of Posterior Information for Attention-Based Speech Recognition

Effective Exploitation of Posterior Information for Attention-Based Speech Recognition

End-to-end attention-based modeling is increasingly popular for tackling sequence-to-sequence mapping tasks. Traditional attention mechanisms utilize prior input information to derive attention, which then conditions the output. However, we believe that knowledge of posterior output information may...

Full description

Bibliographic Details
Main Authors: Jian Tang, Junfeng Hou, Yan Song, Li-Rong Dai, Ian Mcloughlin
Format: Article
Language:English
Published: IEEE 2020-01-01
Series:IEEE Access
Subjects:
Speech recognition
posterior attention
divergence penalty
exposure bias
alternate learning strategy
Online Access:https://ieeexplore.ieee.org/document/9114877/

Internet

https://ieeexplore.ieee.org/document/9114877/

Similar Items