mlsanityData to 7 Oct 2026

Toward Fairness in Speech Recognition: Discovery and mitigation of performance disparities

arXiv 2207.11345 ↗First posted 22 Jul 2022cs.CL cs.SD eess.AScited by 27

Not on the map. Why?

Abstract (arXiv)

As for other forms of AI, speech recognition has recently been examined with respect to performance disparities across different user cohorts. One approach to achieve fairness in speech recognition is to (1) identify speaker cohorts that suffer from subpar performance and (2) apply fairness mitigation measures targeting the cohorts discovered. In this paper, we report on initial findings with both discovery and mitigation of performance disparities using data from a product-scale AI assistant speech recognition system. We compare cohort discovery based on geographic and demographic information to a more scalable method that groups speakers without human labels, using speaker embedding technology. For fairness mitigation, we find that oversampling of underrepresented cohorts, as well as modeling speaker cohort membership by additional input variables, reduces the gap between top- and bottom-performing cohorts, without deteriorating overall recognition accuracy.

Key facts. Cited by 27 papers in the map (data through 7 Oct 2026). Builds on 14 of its 25 references that are in the map. First posted on arXiv 22 Jul 2022.

Builds on

14 of 25 references.

References · 14

  1. 2019A Survey on Bias and Fairness in Machine Learningcited by 1,911
  2. 2017Counterfactual Fairnesscited by 802
  3. 2013Speech Recognition with Deep Recurrent Neural Networkscited by 666
  4. 2017Generalized End-to-End Loss for Speaker Verificationcited by 290
  5. 2020Social Biases in NLP Models as Barriers for Persons with Disabilitiescited by 182
  6. 2020Towards Debiasing Sentence Representationscited by 170
  7. 2019Does Object Recognition Work for Everyone?cited by 149
  8. 2020Saving Face: Investigating the Ethical Concerns of Facial Recognition Auditingcited by 89
  9. 2021Quantifying Bias in Automatic Speech Recognitioncited by 50
  10. 2021Towards Measuring Fairness in Speech Recognition: Casual Conversations Dataset Transcriptionscited by 26
  11. 2019Gender Representation in French Broadcast Corpora and Its Impact on ASR Performancecited by 23
  12. 2021Does Robustness Improve Fairness? Approaching Fairness with Word Substitution Robustness Methods for Text Classificationcited by 20
2 more
  1. 2021ASR4REAL: An extended benchmark for speech modelscited by 14
  2. 2017Comparing Human and Machine Errors in Conversational Speech Transcriptioncited by 7

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 2; 2023: 6; 2024: 5; 2025: 8; 2026: 6.

27 papers.

Sort:

By area:

  1. 2024Towards measuring fairness in speech recognition: Fair-Speech datasetcited by 26
  2. 2025OpenAI's Approach to External Red Teaming for AI Models and Systemscited by 26
  3. 2024On the social bias of speech self-supervised modelscited by 16
  4. 2024Spoken Stereoset: On Evaluating Social Bias Toward Speaker in Speech Large Language Modelscited by 15
  5. 2022A Review of Speech-centric Trustworthy Machine Learning: Privacy, Safety, and Fairnesscited by 13
  6. 2023Improving Fairness and Robustness in End-to-End Speech Recognition through unsupervised clusteringcited by 11
  7. 2024A Contrastive Learning Approach to Mitigate Bias in Speech Modelscited by 9
  8. 2023Explaining Speech Classification Models via Word-Level Audio Segments and Paralinguistic Featurescited by 7
  9. 2025Fairness of Automatic Speech Recognition: Looking Through a Philosophical Lenscited by 5
  10. 2022Global Performance Disparities Between English-Language Accents in Automatic Speech Recognitioncited by 4
  11. 2023Improving Speech Recognition for African American English With Audio Classificationcited by 4
  12. 2024DENOASR: Debiasing ASRs through Selective Denoisingcited by 3
15 more
  1. 2025How to Evaluate Automatic Speech Recognition: Comparing Different Performance and Bias Measurescited by 3
  2. 2023Using Data Augmentations and VTLN to Reduce Bias in Dutch End-to-End Speech Recognition Systemscited by 2
  3. 2025WavShape: Information-Theoretic Speech Representation Learning for Fair and Privacy-Aware Audio Processingcited by 2
  4. 2023Cross-utterance ASR Rescoring with Graph-based Label Propagationcited by 1
  5. 2025CTC-DRO: Robust Optimization for Reducing Language Disparities in Speech Recognitioncited by 1
  6. 2025CO-VADA: A Confidence-Oriented Voice Augmentation Debiasing Approach for Fair Speech Emotion Recognitioncited by 1
  7. 2025Fairness in Dysarthric Speech Synthesis: Understanding Intrinsic Bias in Dysarthric Speech Cloning using F5-TTScited by 1
  8. 2026Where Do Self-Supervised Speech Models Become Unfair?cited by 1
  9. 2026Identifying and typifying demographic unfairness in phoneme-level embeddings of self-supervised speech recognition modelscited by 1
  10. 2026Responsible Benchmarking of Fairness for Automatic Speech Recognitioncited by 1
  11. 2023Elucidate Gender Fairness in Singing Voice Transcriptioncited by 0
  12. 2025Towards Fair ASR For Second Language Speakers Using Fairness Prompted Finetuningcited by 0
  13. 2026Mixture-of-Experts with Intermediate CTC Supervision for Accented Speech Recognitioncited by 0
  14. 2026"OK Aura, Be Fair With Me": Demographics-Agnostic Training for Bias Mitigation in Wake-up Word Detectioncited by 0
  15. 2026A Cross Community Agenda for Speech AIcited by 0