Papers

EDA:面向可核查多模态抑郁评估的证据链框架

EDA: Observe, Verify, and Assess for Evidence-Grounded Multimodal Depression Analysis

Why we picked it

做医疗或心理健康 AI 的开发者该看看这篇:EDA 用冻结 7B 模型,靠证据记录加智能体核查就比直接预测 F1 高了近 10 个点,而且每条结论都能回溯到具体时间段的音视频片段,比常见的黑盒多模态打分好解释多了。

研究团队提出 EDA 框架,将多模态抑郁评估拆解为 Observe、Verify、Assess 三个阶段,解决现有方法把长录像压缩成全局行为描述、观察来源难以核查的问题。在 Observe 阶段,基于文献的声学与视觉线索被定位为带时间区间和行为测量值的 Evidence Records;Verify 阶段由冻结的多模态智能体逐条核查媒体证据是否支持观察结论,通过 Support 准入策略保留被正面验证的记录;Assess 阶段结合统计描述与 PHQ-9 引导的文本分析给出症状证据,跨模态不一致时回看原始文本再做选择。

Body

在 D-Vlog 和 LMVD 两个数据集上的实验分别评估线索标注一致性与下游评估效果:AI 线索标注在共享片段上对人类共识标签达到 91.02% 准确率;使用冻结的 7B 骨干模型,EDA Support 相比直接预测在 D-Vlog Leave-Query-Out 与 LMVD 上分别提升 F1 7.61 和 9.88 个百分点。

这项工作把心理健康的 AI 评估从黑盒打分转向可溯源的行为证据链,对医疗 AI 产品的合规审查和临床信任建设有直接参考价值,也示范了冻结骨干模型加证据筛选的轻量化路径。

Original · arXiv cs.AI

EDA: Observe, Verify, and Assess for Evidence-Grounded Multimodal Depression Analysis

Existing multimodal depression assessment methods often summarize long recordings into global behavioral descriptions. This leaves the source and decision relevance of individual observations difficult to inspect. We propose EDA, an evidence-grounded framework organized around Observe, Verify, and Assess. In Observe, literature-informed acoustic and visual cues are localized into source-linked Evidence Records with temporal intervals and behavioral measurements. In Verify, frozen multimodal agents examine whether the cited media support each observation. The simplified Support admission policy retains positively verified records and their review grounds. In Assess, statistical descriptors and temporal events support reference-based behavioral comparison, while PHQ-9-guided text analysis provides complementary symptom evidence. Cross-modal disagreement triggers review of the original text before final selection. Experiments on D-Vlog and LMVD evaluate human-referenced cue agreement and downstream assessment utility separately. AI cue annotations achieve 91.02% accuracy against human consensus labels on shared source segments. With a frozen 7B backbone, EDA Support improves F1 over direct prediction by 7.61 and 9.88 percentage points on D-Vlog Leave-Query-Out and LMVD, respectively. These findings support explicit, source-linked behavioral evidence for inspectable multimodal depression assessment.

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include multimodal, PHQ-9, agents.

Source attribution: curated from arXiv cs.AI (@Changzeng Fu, Ming Shang, Xinyi Zhang, Jiadong Wang, Runhao Zeng, Shiqi Zhao).