| title | Vividh-ASR: A Complexity-Tiered Benchmark and Optimization Dynamics for Robust Indic Speech Recognition |
|---|---|
| authors | Kush Juvekar, Kavya Manohar, Aditya Srinivas Menon, Arghya Bhattacharya, Kumarmanas Nethil |
| date | 2026-05-13 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2605.13087 |
| affiliation | Adalat AI |
| upvotes | 2 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf |
| abstract | Fine-tuning multilingual ASR models like Whisper for low-resource languages often improves read speech but degrades spontaneous audio performance, a phenomenon we term studio-bias. To diagnose this mismatch, we introduce Vividh-ASR, a complexity-stratified benchmark for Hindi and Malayalam across four tiers: studio, broadcast, spontaneous, and synthetic noise. Through a controlled study of learning-rate timing and curriculum ordering, we find that early large parameter updates improve global WER by 12 absolute points, while a hard-to-easy curriculum adds gains for spontaneous speech. These findings motivate reverse multi-stage fine-tuning (R-MFT), a training recipe that enables a parameter-efficient 244M Whisper model to match or exceed conventionally fine-tuned 769M counterparts. Representational analysis via CKA and SVD reveals effective schedules concentrate adaptation in the decoder, preserving the pre-trained encoder's acoustic geometry. We release the benchmark and models. |
cite
@misc{juvekar2026vividh,
title={Vividh-ASR: A Complexity-Tiered Benchmark and Optimization Dynamics for Robust Indic Speech Recognition},
author={Kush Juvekar and Kavya Manohar and Aditya Srinivas Menon and Arghya Bhattacharya and Kumarmanas Nethil},
year={2026},
eprint={2605.13087},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2605.13087}
}