Black-Box Adaptation of ASR for Accented Speech

Khandelwal, Kartik; Jyothi, Preethi; Awasthi, Abhijeet; Sarawagi, Sunita

doi:10.21437/interspeech.2020-3162

Cited by 5 publications

(4 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Khandelwal et al [37] propose a novel coupling of an opensource accent-tuned local model with a commercial ASR model trained on non-accented speech. The authors introduce the "FineMerge" algorithm to enable a commercial ASR system to guide inference of an accent-tuned ASR system at a character level.…”

Section: System Combinationmentioning

confidence: 99%

Accented Speech Recognition: A Survey

Hinsvark,

Delworth,

Del Rio

et al. 2021

Preprint

View full text Add to dashboard Cite

Automatic Speech Recognition (ASR) systems generalize poorly on accented speech. The phonetic and linguistic variability of accents present hard challenges for ASR systems today in both data collection and modeling strategies. The resulting bias in ASR performance across accents comes at a cost to both users and providers of ASR.We present a survey of current promising approaches to accented speech recognition and highlight the key challenges in the space. Approaches mostly focus on single model generalization and accent feature engineering. Among the challenges, lack of a standard benchmark makes research and comparison especially difficult.

show abstract

Section: System Combinationmentioning

confidence: 99%

Accented Speech Recognition: A Survey

Hinsvark,

Delworth,

Del Rio

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…In recent years, cloud-based speech-to-text services from technology companies such as Google and Amazon have become popular, offering flexibility and convenience. However, while these services exhibit impressive performance for general use-cases, they are not optimal for domain-specific vocabulary [12,13,14]. Moreover, these services are provided as black boxes without access to the inner workings of their ASR models, making it impossible to finetune on domain-specific tasks and datasets.…”

Section: Referencementioning

confidence: 99%

ASR Error Detection via Audio-Transcript entailment

Meripo¹,

Konam²

2022

Preprint

View full text Add to dashboard Cite

Despite improved performances of the latest Automatic Speech Recognition (ASR) systems, transcription errors are still unavoidable. These errors can have a considerable impact in critical domains such as healthcare, when used to help with clinical documentation. Therefore, detecting ASR errors is a critical first step in preventing further error propagation to downstream applications. To this end, we propose a novel end-to-end approach for ASR error detection using audio-transcript entailment. To the best of our knowledge, we are the first to frame this problem as an end-to-end entailment task between the audio segment and its corresponding transcript segment. Our intuition is that there should be a bidirectional entailment between audio and transcript when there is no recognition error and vice versa. The proposed model utilizes an acoustic encoder and a linguistic encoder to model the speech and transcript respectively. The encoded representations of both modalities are fused to predict the entailment. Since doctor-patient conversations are used in our experiments, a particular emphasis is placed on medical terms. Our proposed model achieves classification error rates (CER) of 26.2% on all transcription errors and 23% on medical errors specifically, leading to improvements upon a strong baseline by 12% and 15.4%, respectively.

show abstract

“…With the proliferation of open-source speech technology toolkits and cheap(er) cloud computing, some publics may be able to build or modify these technologies without much or any support from governments of corporations (e.g. Masakhane 18 , Khandelwal et al (2020)). Mugar and Gordon (2020) also emphasise that, in their vision, the aim of civic design by and for publics is care rather than innovation, and space for meaningful interaction between people, rather than efficiency.…”

Section: Speech Technology Design As Civic Designmentioning

confidence: 99%

Language technology practitioners as language managers: arbitrating data bias and predictive bias in ASR

Markl¹,

McNulty²

2022

Preprint

View full text Add to dashboard Cite

Despite the fact that variation is a fundamental characteristic of natural language, automatic speech recognition systems perform systematically worse on non-standardised and marginalised language varieties. In this paper we use the lens of language policy to analyse how current practices in training and testing ASR systems in industry lead to the data bias giving rise to these systematic error differences. We believe that this is a useful perspective for speech and language technology practitioners to understand the origins and harms of algorithmic bias, and how they can mitigate it. We also propose a re-framing of language resources as (public) infrastructure which should not solely be designed for markets, but for, and with meaningful cooperation of, speech communities.

show abstract

Black-Box Adaptation of ASR for Accented Speech

Cited by 5 publications

References 19 publications

Accented Speech Recognition: A Survey

Accented Speech Recognition: A Survey

ASR Error Detection via Audio-Transcript entailment

Language technology practitioners as language managers: arbitrating data bias and predictive bias in ASR

Contact Info

Product

Resources

About