Preserve to Expand: Source-Aware Adaptation and Model Merging for Multilingual LLM-Based ASR

Yuchen Zhang, Haralambos Mouratidis, Ravi Shekhar

YAYIN:26 Eki 2026 ALAN:cs.CL

Özet

Expanding multilingual ASR systems to new languages is challenging when adaptation must preserve previously supported languages without replaying source-language training data. We study this problem in connector-based Speech LLMs, where a trainable connector maps speech representations into a frozen large language model. We evaluate target-update construction and merging strategies for adapting the connector while preserving source-language performance. In single-target expansion, target-only fine-tuning and standard TIES merging learn new languages but cause severe source-language forgetting, while SSU-Wanda+TIES gives the best source-preserving trade-off. In multi-target expansion, update merging provides stable source retention, whereas recursive expansion can improve target adaptation under favorable orders but is highly order-sensitive. These results show that multilingual ASR expansion requires balancing target acquisition, source retention, and interactions among target-language updates.

Özetle: Expanding multilingual ASR systems to new languages is challenging when adaptation must preserve previously supported languages without replaying source-language training data.

Özet

Orijinal Özet (İngilizce)

Expanding multilingual ASR systems to new languages is challenging when adaptation must preserve previously supported languages without replaying source-language training data. We study this problem in connector-based Speech LLMs, where a trainable connector maps speech representations into a frozen large language model. We evaluate target-update construction and merging strategies for adapting the connector while preserving source-language performance. In single-target expansion, target-only fine-tuning and standard TIES merging learn new languages but cause severe source-language forgetting, while SSU-Wanda+TIES gives the best source-preserving trade-off. In multi-target expansion, update merging provides stable source retention, whereas recursive expansion can improve target adaptation under favorable orders but is highly order-sensitive. These results show that multilingual ASR expansion requires balancing target acquisition, source retention, and interactions among target-language updates.

Kaynak: arXiv:2610.35846 · PDF

BibTeX

@article{zhang2026preserve,
  title   = {Preserve to Expand: Source-Aware Adaptation and Model Merging for Multilingual LLM-Based ASR},
  author  = {Yuchen Zhang and Haralambos Mouratidis and Ravi Shekhar},
  journal = {arXiv preprint arXiv:2610.35846},
  year    = {2026},
  url     = {https://arxiv.org/abs/2610.35846}
}

Tartışma

Bu habere emoji ile tepki ver

İptal

Hizli:

Henüz yorum yok. İlk yorumu siz yapın!

Yapıcı ve saygılı yorumlar bekliyoruz. Topluluk kuralları