The Rwth Asr System for Ted-Lium Release 2: Improving Hybrid Hmm With Specaugment

• Published in: ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) pp. 7839 - 7843
• Main Authors: Zhou, Wei, Michel, Wilfried, Irie, Kazuki, Kitza, Markus, Schluter, Ralf, Ney, Hermann
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-05-2020
• Subjects: Acoustics; Hidden Markov models; hybrid BLSTM-HMM; Pipelines; Signal processing; SpecAugment; Speech processing; Speech recognition; TED-LIUM release 2; Training
• ISSN: 2379-190X
• DOI: 10.1109/ICASSP40776.2020.9053573

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.