Evaluating Time-Series Foundation Models and Multimodal Dietary Context for CGM Forecasting

📄 arXiv: 2609.11872v1 📥 PDF

作者: Bowen Zhang, Hsiu-Wen Cheng, Hongyu Yang, Evie L. Shen, Joleen Vansomphone, Yuna Li, Kerry Zhou, Zitian Qu, Suning Zhao, Xiangning Deng, Hua Zhou, Jin J. Zhou

分类: stat.ML, cs.LG

发布日期: 2026-09-10


💡 一句话要点

提出多模态饮食背景下的CGM预测方法以提升糖尿病管理

🎯 匹配领域: 支柱九:具身大模型 (Embodied Foundation Models)

关键词: 持续血糖监测 时间序列模型 多模态融合 糖尿病管理 预测模型 微调技术 饮食背景 临床应用

📋 核心要点

  1. 现有的时间序列基础模型在CGM预测中未能持续超越特定任务的基线,显示出其在糖尿病管理中的局限性。
  2. 论文提出通过轻量级微调和多模态饮食背景融合的方法,以提升CGM预测的准确性和可靠性。
  3. 实验结果表明,微调后的Chronos-Bolt模型在不同人群中显著降低了RMSE,且多模态融合框架有效提升了预测性能。

📝 摘要(中文)

持续血糖监测(CGM)提供高频率的血糖动态测量,能够实现糖尿病管理中的短期血糖预测。尽管时间序列基础模型在预测能力上表现出色,但其在CGM预测中的有效性及多模态饮食背景的附加价值尚不明确。本文通过对八个公共CGM数据集的实证研究,发现零-shot基础模型未能持续超越强大的任务特定基线,而轻量级微调显著提升了预测性能。此外,利用CGMacros评估多模态饮食背景,结果显示融合框架在CGM-only基线基础上减少了约3%的整体RMSE和约15%的餐后RMSE。这些发现表明,基础模型需要针对CGM进行特定适配,饮食背景在餐后期间提供了临床上有意义的信号。

🔬 方法详解

问题定义:本文旨在解决时间序列基础模型在持续血糖监测(CGM)预测中的有效性不足及多模态饮食背景的价值未知的问题。现有方法在糖尿病管理中表现不佳,尤其是在餐后血糖预测方面。

核心思路:论文的核心思路是通过轻量级微调和结合多模态饮食信息来提升CGM预测的准确性。通过对基础模型进行特定任务的适配,能够更好地捕捉血糖动态变化。

技术框架:整体架构包括数据预处理、模型训练和评估三个主要阶段。首先,利用CGMacros提供的多模态数据进行训练,然后通过微调基础模型以适应CGM预测任务,最后评估模型在不同测试集上的表现。

关键创新:最重要的技术创新在于提出了基于残差的融合框架,该框架有效整合了CGM信号、食物图像和营养记录,显著提升了预测性能。与现有方法相比,该框架能够更好地捕捉餐后血糖变化模式。

关键设计:在模型设计中,采用了Chronos-Bolt作为基础模型,并通过微调优化其参数设置。损失函数选择了适合回归任务的均方误差(MSE),并在模型训练中引入了多模态数据以增强模型的泛化能力。

🖼️ 关键图片

fig_0
fig_1
fig_2

📊 实验亮点

实验结果显示,微调后的Chronos-Bolt模型在T1D人群中RMSE降低了6.5%-18.4%,在非糖尿病/T2D人群中降低了8.6%-18.2%。此外,基于残差的融合框架使整体RMSE减少约3%,餐后RMSE减少约15%,显示出显著的性能提升。

🎯 应用场景

该研究的潜在应用领域包括糖尿病患者的个性化管理和饮食建议。通过准确预测血糖变化,患者可以更好地调整饮食和药物使用,从而改善健康状况。未来,该方法有望推广至其他慢性疾病的管理中,提升整体医疗效果。

📄 摘要(原文)

Continuous glucose monitoring (CGM) provides high-frequency measurements of glucose dynamics and enables short-term glucose forecasting for diabetes management. Although time-series foundation models have shown strong general forecasting ability, their effectiveness for CGM prediction and the added value of multimodal dietary context remain unclear. We conduct a comprehensive empirical study using eight public CGM datasets spanning Type 1 diabetes, Type 2 diabetes, and non-diabetes populations. Under a unified protocol across multiple context lengths and prediction horizons, zero-shot foundation models did not consistently outperform strong task-specific baselines such as Elastic Net and PatchTST. In contrast, lightweight fine-tuning substantially improved forecasting performance. For example, fine-tuned Chronos-Bolt reduced RMSE by 6.5%-18.4% in the T1D cohort and by 8.6%-18.2% in the non-diabetes/T2D cohort, with comparable improvements in both in-distribution and out-of-distribution test settings. We further evaluate multimodal dietary context using CGMacros, which provides temporally aligned CGM signals, food images, and macronutrient records. A residual-based fusion framework reduced overall RMSE by approximately 3% and postprandial RMSE by approximately 15% relative to the CGM-only baseline. Moreover, Chronos-based CGM representations were more strongly correlated with observed postprandial glucose increments than representations from LSTM and CatBoost, even after those models incorporated additional dietary modalities, suggesting that pretrained temporal representations better preserve meal-induced excursion patterns. These findings show that foundation models require CGM-specific adaptation for reliable forecasting and that dietary context provides clinically meaningful signals beyond CGM alone, especially during postprandial periods.