Advancing Connectionist Temporal Classification With Attention Modeling
arXiv:1803.05563
Abstract
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.
Accepted at ICASSP 2018
References in corpus (6)
- Exploring Neural Transducers for End-to-End Speech Recognition
- Neural Speech Recognizer: Acoustic-to-Word LSTM Model for Large Vocabulary Speech Recognition
- Gram-CTC: Automatic Unit Selection and Target Decomposition for Sequence Labelling
- Direct Acoustics-to-Word Models for English Conversational Speech Recognition
- Advancing Acoustic-to-Word CTC Model
- Improving the Performance of Online Neural Transducer Models