arXiv：cs.AI· Tony Alex, Wish Suharitdamrong, Sara Atito, Armin Mustafa, Muhammad Awais, Philip J. B. Jackson, Jiankang Deng, Ismail Elezi·· 6 小时前
AudioChaps：用 GRPO 和 CoT 对齐 LALM 实现音频章节化
Listen, Reason, and Segment: Aligning LALMs with Editorial Judgment for Media Chapterization
AI 导读
研究者提出 AudioChaps，通过 GRPO 结合 CoT 推理对端到端 LALM 进行后训练，用于将连续音频流分割为语义连贯的章节。不做 SFT 冷启动的 AudioChaps-R1-Zero 平均 F1 比 SOTA 模型 Audio-Flamingo-3-Think 提升 33 分，完整框架 AudioChaps-R1 提升 49 分。
正文
Abstract:Large Audio Language Models (LALMs) have made rapid progress on standardized benchmarks, yet their deployment in practical media workflows, curation, archival indexing, and content distribution remains largely unrealized. We identify automated audio chapterization, the task of segmenting continuous audio streams into thematically coherent chapters, as a demanding and commercially consequential setting that exposes this gap. Chapterization is challenging because boundaries are defined less by objective acoustic events than by subjective editorial judgment, requiring models to reason sequentially over long acoustic contexts and approximate creator-authored boundary decisions. We present AudioChaps, a post-training framework for aligning end-to-end LALMs for this task via Group Relative Policy Optimization (GRPO) guided by Chain-of-Thought (CoT) reasoning. To support training and evaluation, we curate three datasets: AudioChaps-Alignment, derived from creator-annotated chapter boundaries on YouTube; AudioChaps-CoT, which provides structured supervision for well-formatted, high-quality, and evidence-grounded boundary reasoning; and AudioChaps-Eval, a held-out benchmark for audio chapterization. Applying GRPO directly without a Supervised Fine-Tuning (SFT) cold start, AudioChaps-R1-Zero already improves average F1 by 33 points over the state-of-the-art LALM Audio-Flamingo-3-Think. The AudioChaps framework produces our final aligned LALM, AudioChaps-R1, which improves average F1 by 49 points. These results demonstrate that GRPO-trained LALMs can reliably transform unstructured auditory streams into navigable, structured media. Our code, models, and dataset resources will be released upon acceptance at this https URL.
| Comments: | 19 pages, 9 figures, 8 tables |
| Subjects: | Sound (cs.SD); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Audio and Speech Processing (eess.AS) |
| Cite as: | arXiv:2608.16539 [cs.SD] |
| (or arXiv:2608.16539v2 [cs.SD] for this version) | |
| https://doi.org/10.48550/arXiv.2608.16539 arXiv-issued DOI via DataCite |
Submission history
From: Tony Alex [view email]
[v1]
Mon, 17 Aug 2026 13:14:01 UTC (2,546 KB)
[v2]
Sun, 11 Oct 2026 18:37:15 UTC (2,546 KB)
来源：arXiv：cs.AI · arxiv.org