The RWTH ASR System for TED-LIUM Release 2: Improving Hybrid HMM with SpecAugment
This work provides incremental improvements to ASR systems for speech recognition tasks, specifically on the TED-LIUM dataset.
The paper tackled improving hybrid HMM-based automatic speech recognition on the TED-LIUM corpus by applying SpecAugment data augmentation and fine-tuning, resulting in a 5.6% word error rate that outperforms the previous state-of-the-art by 27% relative.
We present a complete training pipeline to build a state-of-the-art hybrid HMM-based ASR system on the 2nd release of the TED-LIUM corpus. Data augmentation using SpecAugment is successfully applied to improve performance on top of our best SAT model using i-vectors. By investigating the effect of different maskings, we achieve improvements from SpecAugment on hybrid HMM models without increasing model size and training time. A subsequent sMBR training is applied to fine-tune the final acoustic model, and both LSTM and Transformer language models are trained and evaluated. Our best system achieves a 5.6% WER on the test set, which outperforms the previous state-of-the-art by 27% relative.