标签
Introduces CANOE, a multi-agent neuro-symbolic framework for open-ended care plan coordination that uses argumentative computation and human-in-the-loop contestation to improve transparency, safety, and clinical correctness.
本文介绍了CoPlan,一种用于人机协同护理规划的协同智能且可争议的界面,它利用多智能体工作流生成候选干预措施和论点,使人类护理规划者能够在最终计划生成之前检查、质疑和修改建议。该方法在居家养老场景中进行了演示,并为可信赖的人机协同护理规划提供了设计框架。