Sorry, you need to enable JavaScript to visit this website.

ADVANCING CONNECTIONIST TEMPORAL CLASSIFICATION WITH ATTENTION MODELING

Citation Author(s):
Amit Das, Jinyu Li, Rui Zhao, Yifan Gong
Submitted by:
Jinyu Li
Last updated:
13 April 2018 - 7:29pm
Document Type:
Presentation Slides
Document Year:
2018
Event:
Presenters:
Jinyu Li
Paper Code:
SP-L1.5
 

In this study, we propose advancing all-neural speech recognition by directly incorporating attention modeling within the Connectionist Temporal Classification (CTC) framework. In particular, we derive new context vectors using time convolution features to model attention as part of the CTC network. To further improve attention modeling, we utilize content information extracted from a network representing an implicit language model. Finally, we introduce vector based attention weights that are applied on context vectors across both time and their individual components. We evaluate our system on a 3400 hours Microsoft Cortana voice assistant task and demonstrate that our proposed model consistently outperforms the baseline model achieving about 20% relative reduction in word error rates.

up
0 users have voted: