Connecting the Dots: Enduring Performance Gaps in ASR Systems Between Standard and Minority English Varieties
cs.CL
Submitted: 2026-10-13
Updated: 2026-10-13
Comments: Accepted to IEEE SLT 2026. 11 pages including appendix, 1 figure, 6 tables. Code and data: https://github.com/smfsamir/longitudinal-asr-eval
Code: https://github.com/smfsamir/longitudinal-asr-eval
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sources
- Scaling Laws for Neural Language Models
- The Actuary's Final Word on Algorithmic Decision Making
- Less is More: Accurate Speech Recognition & Translation without Web-Scale Data
- Qwen2-Audio Technical Report
- Robust Speech Recognition via Large-Scale Weak Supervision
- Bringing the People Back In: Contesting Benchmark Machine Learning Datasets
- WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing
- Scaling Speech Technology to 1,000+ Languages
- Reproducing Whisper-Style Training Using an Open-Source Toolkit and Publicly Available Data
- Seamless: Multilingual Expressive and Streaming Speech Translation
- Google USM: Scaling Automatic Speech Recognition Beyond 100 Languages
- Omnilingual ASR: Open-Source Multilingual Speech Recognition for 1600+ Languages
- SUPERB: Speech processing Universal PERformance Benchmark
- Open ASR Leaderboard: Towards Reproducible and Transparent Multilingual and Long-Form Speech Recognition Evaluation
- SpeechStew: Simply Mix All Available Speech Recognition Data to Train One Large Neural Network
- MLS: A Large-Scale Multilingual Dataset for Speech Research
- Data Science at the Singularity
- Distributionally Robust Neural Networks for Group Shifts: On the Importance of Regularization for Worst-Case Generalization
- Earnings-21: A Practical Benchmark for ASR in the Wild
- Earnings-22: A Practical Benchmark for Accents in the Wild
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering