arXiv cs.AI· Shakson Isaac, Chirag Patel·· 13 小时前AI 评分26
SSMCGM-Stream：面向个性化糖尿病管理的流式状态空间模型
SSMCGM-Stream: A Streaming State-Space Model for Personalized Diabetes Management
AI 导读
SSMCGM-Stream 是一种流式状态空间模型，随每条血糖读数更新个性化隐藏状态，实现常数时间预测、生理结构化的反事实推理与基于隐藏状态归因的可解释预测。在 T1DEXI 队列上，其预测性能与强神经基线相当，同时大幅降低计算成本，支持设备端持续糖尿病决策支持。基于该模型，SSMCGM-Agent 通过受约束的语言模型接口将预测、解释与反事实转化为有依据的风险感知自然语言建议。
正文 · 原文
Abstract:Continuous glucose monitoring supports diabetes management but requires forecasting models that are accurate, interpretable, and efficient enough for real-time deployment. We present SSMCGM-Stream, a streaming state-space model that maintains a personalized hidden state updated with each incoming glucose reading to provide constant-time forecasting, physiologically structured counterfactual reasoning, and interpretable forecasts through hidden-state attribution. Built on this foundation, SSMCGM-Agent delivers risk-aware, natural-language decision support by translating forecasts, explanations, and counterfactuals into grounded recommendations through a guarded language-model interface. Evaluated on the T1DEXI cohort, SSMCGM-Stream achieves forecasting performance comparable to strong neural baselines while substantially reducing computational cost, enabling continuous on-device diabetes decision support.
| Comments: | Accepted to AI4Science and GenAI4Health NeurIPS 2026 workshops |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.28976 [cs.LG] |
| (or arXiv:2610.28976v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.28976 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Shakson Isaac [view email]
[v1]
Tue, 20 Oct 2026 22:48:57 UTC (1,316 KB)
来源：arXiv cs.AI · arxiv.org