---
格式版本: 2
标题: "GARLIC: Graph Attention-based Relational Learning of Multivariate Time Series in Intensive Care"
原文链接: "https://arxiv.org/abs/2608.10969"
发布日期: "2026-08-11"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:local:strict_original_body"
发布时间证据: "**\\[v1\\]** Tue, 11 Aug 2026 14:29:48 UTC (5,974 KB)"
发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body"
发布时间校准置信度: "high"
发布时间候选数量: 18
发布时间严格候选数量: 6
发布时间原页读取状态: "source template page reused from URL open"
发布时间未找到原因: ""
发布时间校准时间: "2026-08-12T11:20:48+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-08-12T11:16:29+08:00"
入库时间: "2026-08-12T03:20:49.160Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=performance&searchtype=all"
匹配关键词:
  - "performance"
相关厂家:
  []
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 5
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "该论文关于ICU时间序列预测，与超节点、AI Rack、机柜级AI基础设施等完全无关，属于明显无关内容。"
AI质检模型: "ali-deepseek-v4-flash"
AI质检时间: "2026-08-12T11:20:53+08:00"
AI主题相关性: 0
AI来源权威性: 3
AI新颖性: 1
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 1
采集批次: "2026年8月12日2点04分20秒"
采集批次ID: "20260812-020420-289"
去重键: "https://arxiv.org/abs/2608.10969"
---

## Computer Science > Machine Learning

## Title:GARLIC: Graph Attention-based Relational Learning of Multivariate Time Series in Intensive Care

Authors:[Ruirui Wang](https://arxiv.org/search/cs?searchtype=author&query=Wang,+R), [Yanke Li](https://arxiv.org/search/cs?searchtype=author&query=Li,+Y), [Manuel Günther](https://arxiv.org/search/cs?searchtype=author&query=G%C3%BCnther,+M), [Diego Paez-Granados](https://arxiv.org/search/cs?searchtype=author&query=Paez-Granados,+D)

[View PDF](https://arxiv.org/pdf/2608.10969) [HTML (experimental)](https://arxiv.org/html/2608.10969v1)

> Abstract:Healthcare data, such as Intensive Care Unit (ICU) records, comprise heterogeneous multivariate time series sampled at irregular intervals with pervasive missingness. However, clinical applications demand predictive models that are both accurate and interpretable. We present our Graph Attention-based Relational Learning for Intensive Care (GARLIC) model, a novel neural network architecture that imputes missing data through a learnable exponential-decay encoder, captures inter-sensor dependencies via time-lagged summary graphs, and fuses global patterns with cross-dimensional sequential attention. All attention weights and graph edges are learned end-to-end to serve as built-in observation-, signal-, and edge-level explanations. To reconcile auxiliary reconstruction and primary classification objectives, we developed an alternating decoupled optimization scheme that stabilizes training. On three ICU benchmarks (PhysioNet 2012 & 2019, MIMIC-III), GARLIC sets the new state of the art in outcome prediction, significantly improving AUROC and AUPRC over best-performing baselines at comparable computational cost. Ablation studies confirm the contribution of each module, and feature-removal trials validate the fidelity of importance attribution through a monotonic performance drop (full > top 50% > random 50% > bottom 50%). Real-time case studies demonstrate actionable risk warnings with transparent explanations, marking a significant advance toward accurate, explainable deep learning for irregularly sampled ICU time series data. Moreover, we demonstrated \\proposed{}'s superiority in data imputation and classification on various time-series datasets beyond the ICU domain, showing its generalizability and applicability to broader tasks.

| Comments: |  |
| --- | --- |
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | [arXiv:2608.10969](https://arxiv.org/abs/2608.10969) \[cs.LG\] |
|  | (or [arXiv:2608.10969v1](https://arxiv.org/abs/2608.10969v1) \[cs.LG\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2608.10969](https://doi.org/10.48550/arXiv.2608.10969) |

## Submission history

From: Yanke Li \[[view email](https://arxiv.org/show-email/67b20235/2608.10969)\]  
**\[v1\]** Tue, 11 Aug 2026 14:29:48 UTC (5,974 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2608.10969) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))