Model Releases
Sequential Adapter Stacking for Cross-Lingual Low-Resource ASR
arXiv:2609.15758v1 Announce Type: new Abstract: Extending large-scale multilingual automatic speech recognition (ASR) models to low-resource languages remains challenging. Model performance is skewed
arXiv:2609.15758v1 Announce Type: new Abstract: Extending large-scale multilingual automatic speech recognition (ASR) models to low-resource languages remains challenging. Model performance is skewed toward high-resource languages and degrades sharply for languages with limited labeled data and pre-training exposure. To address this, we investigate parameter-efficient approaches for transferring knowledge from resource-rich source languages to low-resource target languages on Whisper. Alongside warm initialization and attention-based fusion, we propose Sequential Adapter Stacking, which places a trainable target-language adapter on top of a frozen source-language adapter. Under controlled experiments, these approaches are evaluated on three target languages unsupported by Whisper -- Asturian, Assamese, and Xhosa -- using source languages with varying degrees of relatedness. Sequential Adapter Stacking with the closest related source consistently and significantly outperforms full fine-tuning across the three targets, with 5--8% relative WER reductions. These gains largely persist with only one hour of target training data.
Related
- Can Large Language Models Reliably Correct Errors in Low-Resource ASR? A Contamination-Aware Case Study on West Frisian
- MoLGE: Mixture of Language Group Experts for Efficient Scaling of Massively Multilingual Speech Recognition
- Sometin Beta Pass Notin (SBPN): Improving Multilingual ASR for Nigerian Languages via Knowledge Distillation
Source: arXiv cs.CL | 2026-09-15