生成用于长文本音频摘要的合成医生-患者对话
Generating Synthetic Doctor-Patient Conversations for Long-form Audio Summarization
AI总结：
本文提出生成合成医生-患者对话数据的方法，用于长文本音频摘要任务，通过多阶段流程生成包含音频和参考笔记的数据集，评估显示级联方法优于端到端模型。
摘要：
长上下文音频推理在训练数据和评估方面存在不足。现有基准针对短上下文任务，而与长上下文推理最相关的开放生成任务在自动评估中面临已知挑战。我们提出一个合成数据生成流程，既作为训练资源又作为受控评估环境，实例化为首次就诊医生-患者对话，任务为SOAP笔记生成。该流程有三个阶段：基于人物的对话生成、多说话者音频合成（含重叠/暂停建模）、房间声学和声音事件，以及基于LLM的参考SOAP笔记生成，完全基于开源模型。我们发布了8,800个合成对话，包含1.3k小时对应的音频和参考笔记。评估当前开源系统时，发现级联方法仍显著优于端到端模型。
英文摘要：
Long-context audio reasoning is underserved in both training data and evaluation. Existing benchmarks target short-context tasks, and the open-ended generation tasks most relevant to long-context reasoning pose well-known challenges for automatic evaluation. We propose a synthetic data generation pipeline designed to serve both as a training resource and as a controlled evaluation environment, and instantiate it for first-visit doctor-patient conversations with SOAP note generation as the task. The pipeline has three stages, persona-driven dialogue generation, multi-speaker audio synthesis with overlap/pause modeling, room acoustics, and sound events, and LLM-based reference SOAP note production, built entirely on open-weight models. We release 8,800 synthetic conversations with 1.3k hours of corresponding audio and reference notes. Evaluating current open-weight systems, we find that cascaded approaches still substantially outperform end-to-end models.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
