Comparison of Automatic Speech Recognition Systems

Kim, Joshua Y.; Liu, Chunfeng; Calvo, Rafael A.; McCabe, Kathryn; Taylor, Silas; Schuller, Björn W.; Wu, Kaihang

doi:10.1007/978-981-19-5538-9_8

Cited by 5 publications

(8 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The optimistic interpretation regarding accuracy would be that ASR performance is nowadays performing well in different locations and social contexts. This is in line with current ASR research that reports sinking WER rates in various conditions and languages due to training effects (Kim et al 2019;Proksch, Wratil, and Wäckerle 2019;Tancoigne et al 2022). The more pessimistic interpretation would be that situations might be more complex than assumed.…”

Section: Discussionsupporting

confidence: 88%

“…Although this is contrary to previous research (Garnerin, Rossato, and Besacier 2003;Pellegrini et al 2012), this is good news because we did not find evidence for bias in the ASR transcriptions for these subgroups. The likely explanation for our null findings is that the performance of ASR systems improved since previous research has been conducted (Kim et al 2019).…”

Section: Discussionmentioning

confidence: 76%

“…The accuracy of ASR transcriptions might be low due to longer, shorter, missing, added text, or compound words (Errattahi et al 2019;Ghannay, Estève, and Camelin 2020). A common measure of ASR accuracy is the word error rate (WER) which is the number of transcription errors divided by the answer length (Kim et al 2019;Tancoigne et al 2022). The higher the WER, the worse the transcription.…”

Section: Introductionmentioning

confidence: 99%

“…Mobile respondents are more used to voice-recordings than PC or tablet users, which improves ASR performance (Revilla and Couper 2021;Rodrigues et al 2019). Previous research also indicates that ASR performance decreases with increasing distance between the microphone and the speaker (Renals and Swietojanski 2017;Kim et al 2019). PC and laptops are therefore potentially more challenging for ASR systems than mobiles due to the need for distant speech recognition.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Keep the noise down: On the performance of automatic speech recognition of voice-recordings in web surveys

Meitinger,

van der Sluis,

Schonlau

2024

Surv Pract

View full text Add to dashboard Cite

Voice-recordings are increasingly implemented in web surveys, but the resulting audio data need to be transcribed before analysis. Since manual coding is too time- and work-intensive, researchers often rely on automatic speech recognition (ASR) systems for the transcription of the voice-recordings. However, ASR tools might create partly incorrect transcriptions and potentially change the content of responses. If the ASR performance (i.e., accuracy and validity) differs by subgroup and contextual factors, a bias is introduced in the analysis of open-ended questions. We assessed the impact of sociodemographic and contextual factors on the accuracy and validity of ASR transcriptions with data from the Longitudinal Internet Studies for the Social Sciences (LISS) panel collected in December 2020. We find that background noise reduces the accuracy and validity of ASR transcriptions. In addition, validity improved when the respondent was alone during the survey. Fortunately, we did not find any evidence of systematic differences across subgroups (age, sex, education), devices or respondent location.

show abstract

Section: Discussionsupporting

confidence: 88%

Section: Discussionmentioning

confidence: 76%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Keep the noise down: On the performance of automatic speech recognition of voice-recordings in web surveys

Meitinger,

van der Sluis,

Schonlau

2024

Surv Pract

View full text Add to dashboard Cite

show abstract

“… 3 YouTube transcript service has shown the lowest Word Error rate ( Kim et al, 2019 ). To validate this, two of the researchers independently watched the videos and extracted the expressions.…”

mentioning

confidence: 99%

“Are we in this together?”: embedding social identity detection in drones improves emergency coordination

Kordoni,

Gavidia-Calderon,

Levine

et al. 2023

Front. Psychol.

View full text Add to dashboard Cite

Autonomous systems, such as drones, are critical for emergency mitigation, management, and recovery. They provide situational awareness and deliver communication services which effectively guide emergency responders’ decision making. This combination of technology and people comprises a socio-technical system. Yet, focusing on the use of drone technology as a solely operational tool, underplays its potential to enhance coordination between the different agents involved in mass emergencies, both human and non-human. This paper proposes a new methodological approach that capitalizes on social identity principles to enable this coordination in an evacuation operation. In the proposed approach, an adaptive drone uses sensor data to infer the group membership of the survivors it encounters during the operation. A corpus of 200 interactions of survivors’ talk during real-life emergencies was computationally classified as being indicative of a shared identity or personal/no identity. This classification model, then, informed a game-theoretic model of human-robot interactions. Bayesian Nash Equilibrium analysis determined the predicted behavior for the human agent and the strategy that the drone needs to adopt to help with survivor evacuation. Using linguistic and synthetic data, we show that the identity-adaptive architecture outperformed two non-adaptive architectures in the number of successful evacuations. The identity-adaptive drone can infer which victims are likely to be helped by survivors and where help from emergency teams is needed. This facilitates effective coordination and adaptive performance. This study shows decision-making can be an emergent capacity that arises from the interactions of both human and non-human agents in a socio-technical system.

show abstract

Challenges of Automatic Speech Recognition for medical interviews - research for Polish language

Kuligowska,

Stanusch,

Koniew

2023

Procedia Computer Science

View full text Add to dashboard Cite

Comparison of Automatic Speech Recognition Systems

Cited by 5 publications

References 29 publications

Keep the noise down: On the performance of automatic speech recognition of voice-recordings in web surveys

Keep the noise down: On the performance of automatic speech recognition of voice-recordings in web surveys

“Are we in this together?”: embedding social identity detection in drones improves emergency coordination

Challenges of Automatic Speech Recognition for medical interviews - research for Polish language

Contact Info

Product

Resources

About