arXiv · cs.AI · Junwei Wu, Runze Yan, Yuzhang Xie, Hanqi Luo, Ziwei Dong, Jiaying Lu, Xiao Hu, Carl Yang2026-11-02 13:00 · 11 小时前模型评分51

Noura 研究：把个人饮食要求变成规划监督信号

Can Personal Requirements Become Supervision for Planning?

导读

论文提出 Noura，把公开临床指南和个人健康档案转成可执行规格，在规划中约束食物选择、份量优化和多日计划校验，并用执行结果与修订记录监督紧凑模型。

分类置信度

中 · 待人工复核

来源级别：媒体与个人

应用领域：医疗与生命科学

事件类型：研究发布

证据摘录

“We present Noura, which pairs LLM planning with executable personal specifications instantiated from public clinical guidance and individual health and dietary profiles.”
“On 4,993 held-out participants, Noura-SFT raises Qwen3-14B's complete-plan pass rate from 24.0% to 72.5% within the same execution loop, cuts rule violations by 97.5% with no observed critical failures, and improves dietary quality.”
“Across four Qwen, Phi, and Llama backbones, self-training alone reduces rule violations by 41-61% on a 1,249-person population subset.”

来源：arXiv · cs.AI · arxiv.org