2025/06/04 by Zheng-Xin Yong, Vineel Pratap, Yong, Zheng-Xin +5
Computer Science · Psychology · #Audio and Speech Processing (eess.AS) #Computation and Language (cs.CL) #FOS: Computer and information sciences #FOS: Electrical engineering #ICT in Developing Communities #Phonetics and Phonology Research #Sound (cs.SD) #Speech Recognition and Synthesis #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2506.04364
openalex publication_date 2025/06/04 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
To build an automatic speech recognition (ASR) system that can serve everyone in the world, the ASR needs to be robust to a wide range of accents including unseen accents. We systematically study how three different variables in training data -- the number of speakers, the audio duration per each individual speaker, and the diversity of accents -- affect ASR robustness towards unseen accents in a low-resource training regime. We observe that for a fixed number of ASR training hours, it is more beneficial to increase the number of speakers (which means each speaker contributes less) than the number of hours contributed per speaker. We also observe that more speakers enables ASR performance gains from scaling number of hours. Surprisingly, we observe minimal benefits to prioritizing speakers with different accents when the number of speakers is controlled. Our work suggests that practitioners should prioritize increasing the speaker count in ASR training data composition for new languages.