根据音频难易程度动态分配计算资源，提升零样本呼吸音分类准确率
Adaptive Test-Time Scaling for Zero-Shot Respiratory Audio Classification
- 分三级推理流程：从快速匹配到深度语言模型分析，按需调用算力
- 9个任务平均AUROC达0.744，近半样本在最便宜层级完成预测
- 疑难样本提升19%性能，简单样本几乎无额外开销，适合医疗筛查场景
自动化呼吸音分析有望实现可扩展、非侵入式的疾病筛查，但受限于标注数据稀缺和专家标注成本高昂。零样本推理无需任务特定监督，但现有方法对所有输入采用统一计算量。本文提出TRIAGE框架，通过分层零样本机制自适应调节测试时计算：将每个音频样本路由至逐步增强的推理阶段——联合音文嵌入空间中的快速标签余弦评分（Tier-L）、基于临床描述符的结构化匹配（Tier-M）以及检索增强的大语言模型推理（Tier-H）。基于置信度的路由模块使简单预测尽早完成，同时将额外计算资源集中于模糊样本，实现约一半样本在最低成本层级退出。在九个呼吸音分类任务上未进行任务特定训练，TRIAGE达到0.744的平均AUROC，优于以往零样本方法，并在多个任务上达到或超过有监督基线表现。分析显示，测试时计算扩容集中在关键区域：不确定样本性能提升最高达19%相对增益，而高置信预测基本不变，成本极低。
原文摘要 · Abstract (English)
Automated respiratory audio analysis promises scalable, non-invasive disease screening, yet progress is limited by scarce labeled data and costly expert annotation. Zero-shot inference eliminates task-specific supervision, but existing methods apply uniform computation to every input regardless of difficulty. We introduce TRIAGE, a tiered zero-shot framework that adaptively scales test-time compute by routing each audio sample through progressively richer reasoning stages: fast label-cosine scoring in a joint audio-text embedding space (Tier-L), structured matching with clinician-style descriptors (Tier-M), and retrieval-augmented large language model reasoning (Tier-H). A confidence-based router finalizes easy predictions early while allocating additional computation to ambiguous inputs, enabling nearly half of all samples to exit at the cheapest tier. Across nine respiratory classification tasks without task-specific training, TRIAGE achieves a mean AUROC of 0.744, outperforming prior zero-shot methods and matching or exceeding supervised baselines on multiple tasks. Our analysis show that test-time scaling concentrates gains where they matter: uncertain cases see up to 19% relative improvement while confident predictions remain unchanged at minimal cost.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。