Paper Detail
Ziang Ren, Guodong Lin, Yuchen Ai, Kaize Tan, Wei-Qiang Zhang
Large-scale pretrained ASR models such as Whisper exhibit strong multilingual capabilities. However, fine-tuning on low-resource languages often causes catastrophic forgetting. Although continual learning mitigates this issue, existing methods struggle to regulate cross-task interference in multilingual settings, where dominant languages bias optimization. We propose Unified Gradient Projection (UGP), which constrains parameter updates using reference gradients from language-balanced replay in a unified projection space. By equalizing per-language contributions in the projection, UGP reduces dominant-language bias and improves cross-lingual stability. We further show that combining gradient-level projection with data-level replay yields complementary gains in stability and plasticity. Across diverse low-resource language groups and model scales, UGP enables effective adaptation while substantially mitigating forgetting. On Whisper-large-v3, it achieves near-zero average forgetting.
No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.
No ranking explanation is available yet.
No tags.
@article{ren2026unified,
title = {Unified Gradient Projection: Language-Balanced Continual Learning for Multilingual Low-Resource ASR},
author = {Ziang Ren and Guodong Lin and Yuchen Ai and Kaize Tan and Wei-Qiang Zhang},
year = {2026},
abstract = {Large-scale pretrained ASR models such as Whisper exhibit strong multilingual capabilities. However, fine-tuning on low-resource languages often causes catastrophic forgetting. Although continual learning mitigates this issue, existing methods struggle to regulate cross-task interference in multilingual settings, where dominant languages bias optimization. We propose Unified Gradient Projection (UGP), which constrains parameter updates using reference gradients from language-balanced replay in a},
url = {https://arxiv.org/abs/2607.11163},
keywords = {cs.CL, cs.SD, eess.AS},
eprint = {2607.11163},
archiveprefix = {arXiv},
}
{}