用大模型分析青少年语音转录文本，自动识别自杀风险
In-context learning capabilities of Large Language Models to detect suicide risk among adolescents from speech transcripts
- 用大模型进行上下文学习，通过文本特征判断风险
- 仅用转录文本达到0.68准确率（F1=0.7），排名前四
- 适合心理评估自动化、精神健康技术研究者
青少年早期自杀风险检测至关重要，但现有评估方法面临可扩展性挑战。本文参与首个SpeechWellness挑战（SW1），旨在通过语音分析评估中国青少年的自杀风险。由于语音匿名化限制，研究聚焦语言特征，利用大语言模型（LLMs）对转录文本进行分类。采用DSPy进行系统性提示工程，构建了稳健的上下文学习方法，在语言和声学标记上均优于传统微调。系统在180多个提交中分别获得第三和第四名，仅使用转录文本即实现0.68的准确率（F1=0.7）。消融分析表明，增加提示样本可提升性能（p=0.003），效果因模型类型和规模而异。研究推进了自动化自杀风险评估，展示了大模型在心理健康应用中的价值。
原文摘要 · Abstract (English)
Early suicide risk detection in adolescents is critical yet hindered by scalability challenges of current assessments. This paper presents our approach to the first SpeechWellness Challenge (SW1), which aims to assess suicide risk in Chinese adolescents through speech analysis. Due to speech anonymization constraints, we focused on linguistic features, leveraging Large Language Models (LLMs) for transcript-based classification. Using DSPy for systematic prompt engineering, we developed a robust in-context learning approach that outperformed traditional fine-tuning on both linguistic and acoustic markers. Our systems achieved third and fourth places among 180+ submissions, with 0.68 accuracy (F1=0.7) using only transcripts. Ablation analyses showed that increasing prompt example improved performance (p=0.003), with varying effects across model types and sizes. These findings advance automated suicide risk assessment and demonstrate LLMs' value in mental health applications.
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。