Omni-CLST: Error-aware Curriculum Learning with guided Selective chain-of-Thought for audio question answering
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866912591269330944 |
|---|---|
| author | Zhao, Jinghua Su, Hang Fan, Lichun Luo, Zhenbo Wang, Hui Sun, Haoqin Qin, Yong |
| author_facet | Zhao, Jinghua Su, Hang Fan, Lichun Luo, Zhenbo Wang, Hui Sun, Haoqin Qin, Yong |
| contents | With the rapid progress of large audio-language models (LALMs), audio question answering (AQA) has emerged as a challenging task requiring both fine-grained audio understanding and complex reasoning. While current methods mainly rely on constructing new datasets via captioning or reasoning traces, existing high-quality AQA data remains underutilized. To address this, we propose Omni-CLST, an error-aware Curriculum Learning framework with guided Selective Chain-of-Thought. The framework efficiently leverages existing high-quality dataset through two key strategies: an error-aware curriculum that organizes samples by difficulty, and a guided thought dropout mechanism that focuses reasoning on challenging cases. Experiments show that Omni-CLST achieves 73.80% on MMAU-mini and a new state of the art of 64.30% on MMAR, demonstrating robust generalization in multimodal audio-language understanding. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_12275 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Omni-CLST: Error-aware Curriculum Learning with guided Selective chain-of-Thought for audio question answering Zhao, Jinghua Su, Hang Fan, Lichun Luo, Zhenbo Wang, Hui Sun, Haoqin Qin, Yong Sound Artificial Intelligence Audio and Speech Processing With the rapid progress of large audio-language models (LALMs), audio question answering (AQA) has emerged as a challenging task requiring both fine-grained audio understanding and complex reasoning. While current methods mainly rely on constructing new datasets via captioning or reasoning traces, existing high-quality AQA data remains underutilized. To address this, we propose Omni-CLST, an error-aware Curriculum Learning framework with guided Selective Chain-of-Thought. The framework efficiently leverages existing high-quality dataset through two key strategies: an error-aware curriculum that organizes samples by difficulty, and a guided thought dropout mechanism that focuses reasoning on challenging cases. Experiments show that Omni-CLST achieves 73.80% on MMAU-mini and a new state of the art of 64.30% on MMAR, demonstrating robust generalization in multimodal audio-language understanding. |
| title | Omni-CLST: Error-aware Curriculum Learning with guided Selective chain-of-Thought for audio question answering |
| topic | Sound Artificial Intelligence Audio and Speech Processing |
| url | https://arxiv.org/abs/2509.12275 |