基于异构语音协议的证据受限心理健康推理
Evidence-Bounded Mental Health Reasoning from Heterogeneous Speech Protocols
浏览论文内容
中文总结 AI 辅助
本研究针对现有心理健康筛查模型的证据边界问题，提出协议感知的证据控制框架EviBound，在抑郁症筛查任务中达到0.8658的AUROC，且实现零宣称违规。
中文摘要 AI 辅助
使用多模态语音和文本的计算心理健康筛查已展现出巨大潜力。然而，现有模型通常假设所有临床语音协议具有同等证据效力。现实中，从自由访谈到固定朗读任务的异构协议所提供的证据存在本质差异；强制统一推理会抹平这些边界，导致模型从无关文本中虚构症状或过度宣称支持。即便先进的长思维链大语言模型（LLM）也无法解决该问题，因为自由形式的推理会加剧边界违规。为应对此问题，我们将多模态筛查重新表述为证据受限推理问题。我们推出了证据包基准（Evidence Package Benchmark），整合了来自6个异构来源的1870个包，附带明确的模态掩码和证据权限。我们进一步提出了协议感知的证据控制框架EviBound。与直接的LLM提示不同，EviBound使用配置感知规划器限制推理范围，通过五向声学共识协调证据工具，并强制边界评判器抑制无依据的宣称。实验结果显示，EviBound在保留测试集上的抑郁症AUROC达到0.8658，比最强的直接全模态基准高出0.0811 AUROC，同时保持零宣称违规。我们的工作超越了无约束的准确性，朝着更安全的临床自然语言处理研究所需的证据一致、协议感知系统迈进。
英文摘要
Computational mental health screening using multimodal speech and text has shown great promise. However, existing models often assume all clinical speech protocols carry equivalent evidentiary validity. In reality, heterogeneous protocols, from free interviews to fixed reading tasks, support fundamentally different evidence. Forcing uniform reasoning flattens these boundaries, causing models to hallucinate symptoms from irrelevant text or overclaim support. Even advanced long chain-of-thought LLMs fail to resolve this issue, as free-form reasoning can exacerbate boundary violations. To address this, we reformulate multimodal screening as an evidence-bounded reasoning problem. We introduce the Evidence Package Benchmark, integrating 1,870 packages across six heterogeneous sources with explicit modality masks and evidence permissions. We further propose EviBound, a protocol-aware evidence control framework. Unlike direct LLM prompting, EviBound uses a profile-aware planner to restrict reasoning scope, orchestrates evidence tools via five-way acoustic consensus, and enforces a boundary critic to suppress unsupported claims. Empirical results show EviBound achieves a held-out test Depression AUROC of 0.8658, exceeding the strongest direct omni-modal baseline by +0.0811 AUROC while maintaining zero claim violations. Our work moves beyond unconstrained accuracy toward evidence-consistent, protocol-aware systems for safer clinical NLP research.
发表机构
- Beijing University of Posts and Telecommunications（北京邮电大学）
- Harbin Institute of Technology（哈尔滨工业大学）
- Renyixun Health Technology Co., Ltd.（仁医讯健康科技有限公司）
- GDIIST（广东省智能制造研究所）
- Tencent（腾讯）
机构由 AI 辅助整理，请以论文原文为准。
