An Overview of Noise-Robust Automatic Speech Recognition

By Jinyu Li, Li Deng, Yifan Gong, Reinhold Häb-Umbach. 2014

614 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Speech & Audio, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Speech & Audio

In Signal Processing

Citations in the corpus

Benchmarking Neural Network Robustness to Common Corruptions and Perturbations (2019, background): “Speech recognition research emphasizes robustness to common corruptions rather than worst-case, adversarial corruptions (Li et al., 2014; Mitra et al., 2017).”
Light Gated Recurrent Units for Speech Recognition (2018, background): “This lack of robustness typically happens when facing challenging acoustic conditions [13], characterized by considerable levels of non-stationary noise and acoustic reverberation [14]–[22].”
Deep Learning: Methods and Applications (2014, background): “See a comprehensive review in [220] and some additional review literature or original work in [4, 82, 119, 140, 230, 370, 404, 431, 444].”
Improved cepstra minimum-mean-square-error noise reduction algorithm for robust speech recognition (2017, background): “INTRODUCTION Environment robustness in automatic speech recognition (ASR) remains a difficult problem despite many years of research [1][2][3][4][5][6][7][8].”
Conditional Diffusion Probabilistic Model for Speech Enhancement (2022, background): “Speech enhancement, a key element for immersive audio experiences in telecommunication as well as a crucial front-end processor for robust speech recognition [1, 2], assistive hearing [3], and robust speaker recognition [4, 5], is a challenging and still unsolved problem in audio processing.”
Feature space maximum a posteriori linear regression for adaptation of deep neural networks (2014, background): “A recent survey on these adaptation methods can be found in [9].”
Improving Robustness of Deep Neural Network Acoustic Models via Speech Separation and Joint Adaptive Training (2015, methodology): “…not updated during test time, we call the proposed training strategy ‘adaptive’ to differentiate it from the simpler approach of retraining the acoustic model using the transformed features without modifying the transformation in itself–an approach that has been classi ﬁ ed as joint training [20].”
Bone-conducted speech enhancement using deep denoising autoencoder (2018, background): “In the literature, the limited generalization capability is a common issue in many speech-related applications, such as ASR (Li et al., 2014; Huang et al., 2015), speaker recognition (Kolbœk et al.”
Conditional Teacher-student Learning (2019, methodology): “ASR suffers from performance degradation when a well-trained acoustic model is applied in a new domain [19].”
Continuous Speech Separation with Conformer (2020, neutral): “Thanks to the advance in deep learning, a drastic improvement on accuracy and robustness has been brought to modern automatic speech recognition (ASR) system in the past decade [1, 2, 3, 4, 5, 6, 7, 8].”
Single Channel Target Speaker Extraction and Recognition with Speaker Beam (2018, background): “With the deployment of speech driven home devices, there has been an increased interest for noise robust automatic speech recognition (ASR) [1, 2].”
Cycle-Consistent Speech Enhancement (2018, background): “More importantly, speech enhancement is widely applied as a front-end pre-processing stage to improve the performance of automatic speech recognition (ASR) [2, 3, 4, 5, 6] and speaker recognition under noisy conditions [7, 8].”
Uncertainty propagation through deep neural networks (2015, neutral): “In order to achieve an acceptable level of recognition robustness against these effects, the distorted speech is usually de-noised using speech enhancement algorithms [1,2].”
Deep Learning for Environmentally Robust Speech Recognition (2017, background): “A systematic overview of traditional approaches on the back-end can be found in Li et al. (2014).”
Benchmarking Neural Network Robustness to Common Corruptions and Surface Variations (2018, background): “Speech recognition research emphasizes robustness to common corruptions over worst-case, adversarial corruptions [39, 46].”
Unsupervised adaptation with domain separation networks for robust speech recognition (2017, background): “However, ASR systems still suffer from large performance degradation when acoustic mismatch exists between the training and test conditions [7, 8].”
Neural Network Adaptive Beamforming for Robust Multichannel Speech Recognition (2016, background): “While automatic speech recognition (ASR) performance has improved dramatically in recent years, particularly with the advent of deep learning [5], performance in realistic noisy and far-field scenarios is still far-behind clean speech conditions [6, 7, 8] To improve robustness, microphone arrays are”