Teaching Audio Models to Reason: A Unified Framework for Source- and Layer-wise Distillation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909802094919680 |
|---|---|
| author | Yang, Runyan Si, Yuke Gao, Yingying Feng, Junlan Deng, Chao Zhang, Shilei |
| author_facet | Yang, Runyan Si, Yuke Gao, Yingying Feng, Junlan Deng, Chao Zhang, Shilei |
| contents | While large audio language models excel at tasks like ASR and emotion recognition, they still struggle with complex reasoning due to the modality gap between audio and text as well as the lack of structured intermediate supervision. To address this, we propose a unified knowledge distillation framework to transfer reasoning capabilities from a high-capacity textual teacher model to a student audio models while preserving its acoustic competence. Our method introduces two key dimensions: source-wise distillation, which leverages both textual and acoustic teachers to provide complementary modality-specific supervision; and layer-wise distillation, which aligns teacher signals with appropriate student layers to improve transfer efficiency. This dual-dimensional strategy enables fine-grained control over the distillation process, effectively bridging the gap between symbolic reasoning and speech representations. Experimental results show significant improvements in audio reasoning performance, demonstrating the effectiveness of our framework as a reasoning transfer solution for audio modeling. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_18579 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Teaching Audio Models to Reason: A Unified Framework for Source- and Layer-wise Distillation Yang, Runyan Si, Yuke Gao, Yingying Feng, Junlan Deng, Chao Zhang, Shilei Audio and Speech Processing Computation and Language Sound While large audio language models excel at tasks like ASR and emotion recognition, they still struggle with complex reasoning due to the modality gap between audio and text as well as the lack of structured intermediate supervision. To address this, we propose a unified knowledge distillation framework to transfer reasoning capabilities from a high-capacity textual teacher model to a student audio models while preserving its acoustic competence. Our method introduces two key dimensions: source-wise distillation, which leverages both textual and acoustic teachers to provide complementary modality-specific supervision; and layer-wise distillation, which aligns teacher signals with appropriate student layers to improve transfer efficiency. This dual-dimensional strategy enables fine-grained control over the distillation process, effectively bridging the gap between symbolic reasoning and speech representations. Experimental results show significant improvements in audio reasoning performance, demonstrating the effectiveness of our framework as a reasoning transfer solution for audio modeling. |
| title | Teaching Audio Models to Reason: A Unified Framework for Source- and Layer-wise Distillation |
| topic | Audio and Speech Processing Computation and Language Sound |
| url | https://arxiv.org/abs/2509.18579 |