CohortFlow: A Steerable Clinical Cohort Curation Agent with Localized Semantic Correction

Computer Science > Artificial Intelligence arXiv:2610.36100 (cs) [Submitted on 26 Oct 2026] Title: CohortFlow: A Steerable Clinical Cohort Curation Agent with Localized Semantic Correction Authors: Neil Chen, Adrian Serapio, Hanxue Gu, Evan Lee, Kang Wang, Yang Yang.

计算机科学 > 人工智能 arXiv:2610.36100 (cs) [提交于 2026 年 10 月 26 日] 标题：CohortFlow：一种具有局部语义修正功能的可引导临床队列筛选代理。作者：Neil Chen, Adrian Serapio, Hanxue Gu, Evan Lee, Kang Wang, Yang Yang。

Abstract: Language model agents for clinical cohort curation should accommodate evolving eligibility criteria without repeatedly processing unchanged requirements. We introduce CohortFlow, a steerable agent that represents natural language corrections as localized updates to a structured cohort specification and reconciles them with a persistent execution graph.

摘要：用于临床队列筛选的语言模型代理应能适应不断变化的资格标准，而无需重复处理未更改的需求。我们引入了 CohortFlow，这是一种可引导的代理，它将自然语言修正表示为对结构化队列规范的局部更新，并将其与持久执行图进行协调。

Explicitly recorded semantic dependencies enable affected computations to be selectively invalidated while preserving valid prior results. On 100 scenarios derived from clinical-trial eligibility criteria, comprising 126 correction turns, localized steering achieves 77.0% requirement micro-F1 versus 76.4% for full regeneration using the same 8B model, a small numerical advantage that is not statistically significant ($p=0.493$).

显式记录的语义依赖关系使得受影响的计算能够被选择性地失效，同时保留有效的先前结果。在源自临床试验资格标准的 100 个场景（包含 126 次修正轮次）中，局部引导实现了 77.0% 的需求微观 F1 值，而使用相同 8B 模型进行完全重新生成的结果为 76.4%，这一微小的数值优势在统计学上并不显著 ($p=0.493$)。

It reduces mean correction latency by 64.8% after completion and 44.3% during execution, with 38.1% fewer model calls. In a separate proof-of-concept evaluation over MIMIC, all 13 cohort states execute successfully, while correction trajectories demonstrate selective preservation of unaffected computations and recomputation of changed or dependent nodes.

它在完成后将平均修正延迟降低了 64.8%，在执行期间降低了 44.3%，模型调用次数减少了 38.1%。在针对 MIMIC 的独立概念验证评估中，所有 13 个队列状态均成功执行，同时修正轨迹展示了对未受影响计算的选择性保留，以及对已更改或相关节点的重新计算。

Together, these results show that localized semantic steering can substantially reduce correction latency and model calls while maintaining strong requirement-level accuracy.

总之，这些结果表明，局部语义引导可以在保持强大的需求级准确性的同时，大幅降低修正延迟和模型调用次数。