MetaSICL: Adapting Audiroty LLM via Meta Speech In-Context Learning
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866916046280065024 |
|---|---|
| author | Zheng, Haolong Wang, Siyin Jin, Zengrui Hasegawa-Johnson, Mark |
| author_facet | Zheng, Haolong Wang, Siyin Jin, Zengrui Hasegawa-Johnson, Mark |
| contents | Auditory Large Language Models (LLMs) have demonstrated strong performance across a wide range of speech and audio understanding tasks. Nevertheless, they often struggle when applied to low-resource tasks. In case in-domain labeled data are scarce or mismatched with the true test distribution, direct fine-tuning can be brittle. In-Context Learning (ICL) provides a training-free, inference-time solution by adapting auditory LLMs through conditioning on a few in-domain demonstrations. In this work, we first show that $\textit{Vanilla ICL}$, improves zero-shot performance across diverse speech and audio tasks for selected models which suggest that this ICL adaptation capability can be generalized to multimodal setting. Building on this, we propose $\textbf{Meta Speech In-Context Learning (MetaSICL)}$, a post-training recipe utilizes only high resource speech data from various tasks intending to strengthen model's in-context learning capability. Experiments indicate our proposed method outperforms direct fine-tuning in low-resource scenario. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_18904 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | MetaSICL: Adapting Audiroty LLM via Meta Speech In-Context Learning Zheng, Haolong Wang, Siyin Jin, Zengrui Hasegawa-Johnson, Mark Sound Artificial Intelligence Computation and Language Auditory Large Language Models (LLMs) have demonstrated strong performance across a wide range of speech and audio understanding tasks. Nevertheless, they often struggle when applied to low-resource tasks. In case in-domain labeled data are scarce or mismatched with the true test distribution, direct fine-tuning can be brittle. In-Context Learning (ICL) provides a training-free, inference-time solution by adapting auditory LLMs through conditioning on a few in-domain demonstrations. In this work, we first show that $\textit{Vanilla ICL}$, improves zero-shot performance across diverse speech and audio tasks for selected models which suggest that this ICL adaptation capability can be generalized to multimodal setting. Building on this, we propose $\textbf{Meta Speech In-Context Learning (MetaSICL)}$, a post-training recipe utilizes only high resource speech data from various tasks intending to strengthen model's in-context learning capability. Experiments indicate our proposed method outperforms direct fine-tuning in low-resource scenario. |
| title | MetaSICL: Adapting Audiroty LLM via Meta Speech In-Context Learning |
| topic | Sound Artificial Intelligence Computation and Language |
| url | https://arxiv.org/abs/2601.18904 |