A Conformer Based Acoustic Model for Robust Automatic Speech Recognition
This work addresses robust speech recognition for noisy environments, representing an incremental improvement over existing methods.
This paper tackles robust automatic speech recognition by introducing a Conformer-based acoustic model that replaces the recurrent network in an existing WRBN architecture with a convolution-augmented attention mechanism. The model achieves a 6.25% word error rate on the CHiME-4 corpus, outperforming WRBN by 8.4% relatively while being 18.3% smaller and reducing training time by 79.6%.
This study addresses robust automatic speech recognition (ASR) by introducing a Conformer-based acoustic model. The proposed model builds on the wide residual bi-directional long short-term memory network (WRBN) with utterance-wise dropout and iterative speaker adaptation, but employs a Conformer encoder instead of the recurrent network. The Conformer encoder uses a convolution-augmented attention mechanism for acoustic modeling. The proposed system is evaluated on the monaural ASR task of the CHiME-4 corpus. Coupled with utterance-wise normalization and speaker adaptation, our model achieves $6.25\%$ word error rate, which outperforms WRBN by $8.4\%$ relatively. In addition, the proposed Conformer-based model is $18.3\%$ smaller in model size and reduces total training time by $79.6\%$.