2020/10/23 by Jagadeesh Balam, Balam, Jagadeesh, Jocelyn Huang +9 · 6 citations
Computer Science · Engineering · #Artificial intelligence #Artificial neural network #Audio and Speech Processing (eess.AS) #Computer science #Deep neural networks #End-to-end principle #FOS: Electrical engineering #Machine learning #Music and Audio Processing #Noise (video) #Robustness (evolution) #Schema (genetic algorithms) #Speech Recognition and Synthesis #Speech and Audio Processing #Speech recognition #Training set #eess.AS #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2010.12715
published in arXiv (Cornell University) (Cornell University)
arxiv created 2020/10/23 · openalex publication_date 2020/10/23 · arxiv updated 2020/10/27 · openalex created_date 2020/10/29 · openalex updated_date 2026/07/28
We present our experiments in training robust to noise an end-to-end automatic speech recognition (ASR) model using intensive data augmentation. We explore the efficacy of fine-tuning a pre-trained model to improve noise robustness, and we find it to be a very efficient way to train for various noisy conditions, especially when the conditions in which the model will be used, are unknown. Starting with a model trained on clean data helps establish baseline performance on clean speech. We carefully fine-tune this model to both maintain the performance on clean speech, and improve the model accuracy in noisy conditions. With this schema, we trained robust to noise English and Mandarin ASR models on large public corpora. All described models and training recipes are open sourced in NeMo, a toolkit for conversational AI.