Study of the Performance of Automatic Speech Recognition Systems in Speakers with Parkinson’s Disease

Moro-Velázquez, Laureano; Cho, Jaejin; Watanabe, Shinji; Hasegawa‐Johnson, Mark; Scharenborg, Odette; Kim, Heejin; Dehak, Najim

doi:10.21437/interspeech.2019-2993

Cited by 20 publications

(7 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, ASR for child speech is proven more challenging than that for adult speech, due to children's shorter vocal tracts, slower and more variable speaking rate and inaccurate articulation [8]. A speech impairment is known to cause many problems for standard ASR systems, e.g., for impairments related to dysarthria [9], stroke survival, oral cancer [10] or cleft lip and palate [11]. Additionally, recent studies demonstrate how voice assistants perpetuate a racial divide by misrecognising the speech of black speakers more often than of white speakers [2,7].…”

Section: Introductionmentioning

confidence: 99%

Quantifying Bias in Automatic Speech Recognition

Feng¹,

Kudina²,

Halpern³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Automatic speech recognition (ASR) systems promise to deliver objective interpretation of human speech. Practice and recent evidence suggests that the state-of-the-art (SotA) ASRs struggle with the large variation in speech due to e.g., gender, age, speech impairment, race, and accents. Many factors can cause the bias of an ASR system. Our overarching goal is to uncover bias in ASR systems to work towards proactive bias mitigation in ASR. This paper is a first step towards this goal and systematically quantifies the bias of a Dutch SotA ASR system against gender, age, regional accents and non-native accents. Word error rates are compared, and an in-depth phoneme-level error analysis is conducted to understand where bias is occurring. We primarily focus on bias due to articulation differences in the dataset. Based on our findings, we suggest bias mitigation strategies for ASR development.

show abstract

Section: Introductionmentioning

confidence: 99%

Quantifying Bias in Automatic Speech Recognition

Feng¹,

Kudina²,

Halpern³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Our implementation of the CycleGAN-VC is based on a Py-Torch implementation of CycleGAN-VC by 1 . The CycleGAN-VC is used as the baseline model.…”

Section: Cyclegan-vcmentioning

confidence: 99%

“…Dysarthria can greatly reduce a person's quality of life and independence. Operating home appliances through voice could greatly improve these people's lives; however, dysarthric speech recognition performance is not good enough yet for practical applications, which means that there is a great need for high performance dysarthric speech recognition [1].…”

Section: Introductionmentioning

confidence: 99%

The Effectiveness of Time Stretching for Enhancing Dysarthric Speech for Improved Dysarthric Speech Recognition

Prananta¹,

Halpern²,

Feng³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

In this paper, we investigate several existing and a new stateof-the-art generative adversarial network-based (GAN) voice conversion method for enhancing dysarthric speech for improved dysarthric speech recognition. We compare key components of existing methods as part of a rigorous ablation study to find the most effective solution to improve dysarthric speech recognition. We find that straightforward signal processing methods such as stationary noise removal and vocoder-based time stretching lead to dysarthric speech recognition results comparable to those obtained when using state-of-the-art GANbased voice conversion methods as measured using a phoneme recognition task. Additionally, our proposed solution of a combination of MaskCycleGAN-VC and time stretched enhancement is able to improve the phoneme recognition results for certain dysarthric speakers compared to our time stretched baseline.

show abstract

“…In addition, child speech recognition is also proven to be more difficult than adult speech [51]. Another important challenge for current ASR systems is the speech impairment, i.e., ASR systems perform quite bad on speech spoken by people with dysarthria [52], oral cancer [53], and cleft lip and palate [54]. Last but not least, speech variance in accents and regions has also shown effecting the ASR performance a lot [55,56].…”

Section: Related Work On Bias Researchmentioning

confidence: 99%

Mitigating bias against non-native accents

Zhang¹,

Zhang²,

Halpern³

et al. 2022

Interspeech 2022

Self Cite

View full text Add to dashboard Cite

Automatic Speech Recognition (ASR) systems have seen substantial improvements in the past decade; however, not for all speaker groups. Recent research shows that bias exists against different types of speech, including non-native accents, in state-of-the-art (SOTA) ASR systems. To attain inclusive speech recognition, i.e., ASR for everyone irrespective of how one speaks or the accent one has, bias mitigation is essential and necessary. In this thesis, two SOTA ASR systems (one is based on the recurrent neural network (RNN) and the other is based on the transformer architecture) are built to uncover and quantify the bias against non-native accents. Here I focus on bias mitigation against non-native accents using two different approaches: data augmentation and by using more effective training methods. For data augmentation, an autoencoder-based cross-lingual voice conversion (VC) model is used to increase the amount of non-native accented speech training data in addition to data augmentation through speed perturbation. Moreover, I investigate two training methods, i.e., fine-tuning and Domain Adversarial Training (DAT), to see whether they can utilize the available non-native accented speech data more effectively than a standard training approach. Experimental results show for the transformer-based ASR model: (1) adding VC-generated and speed-perturbed data to train the ASR model gives the best bias mitigation performance and the lowest word error rate (WER); (2) fine-tuning reduces the bias against non-native accents but at the cost of native accent performance; and (3) compared with the standard training method, DAT does not leads to further bias reduction. While for the RNN-based ASR model, all the 4 bias mitigation approaches do not show obvious benefits.

show abstract

Study of the Performance of Automatic Speech Recognition Systems in Speakers with Parkinson’s Disease

Cited by 20 publications

References 29 publications

Quantifying Bias in Automatic Speech Recognition

Quantifying Bias in Automatic Speech Recognition

The Effectiveness of Time Stretching for Enhancing Dysarthric Speech for Improved Dysarthric Speech Recognition

Mitigating bias against non-native accents

Contact Info

Product

Resources

About