Agentic discovery of blood biomarker from distilled private health records
- 类型：arxiv
- 标识：2610.04749
- 链接：https://arxiv.org/abs/2610.04749
- 主分类：agent
- 形态：method
- TLDR：Routine complete blood counts (CBCs) could yield new biomarkers, but the private records needed to evaluate candidates cannot be shared with frontier language model agents that excel at discovery. We distilled the evidence held in the Clalit Health Services panel of over 5.4 million patients into a released scoring tool: for each of 13 immune-mediated diseases, a graph attention network was trained inside the data boundary to predict the case-control AUC of candidate CBC expressions, and only the trained weights were released. The tool grounds an agent's propose-score-refine loop in real-world
- 待LLM分类：否
- 标题中文：从蒸馏后的私有健康记录中以 Agent 方式发现血液生物标志物
- TLDR中文：常规全血细胞计数 (CBC) 可能产生新的生物标志物，但评估候选标志物所需的私有记录无法与擅长发现的前沿语言模型 Agent 共享。我们将 Clalit Health Services 超过 540 万患者的证据蒸馏为一个发布的评分工具：针对 13 种免疫介导疾病，在数据边界内训练图注意力网络以预测候选 CBC 表达式的病例-对照 AUC，仅发布训练后的权重。该工具将 Agent 的"提出—评分—迭代"循环锚定在真实世界……
- 来源文件：
- /inbox/tom/_candidates/2026-10-07-agent-rag-longcontext-candidates.json