[论文解读] Early Recognition of Sepsis with Gaussian Process Temporal Convolutional Networks and Dynamic Time Warping

Michael Moor, Max Horn+3 more|arXiv (Cornell University)|Feb 5, 2019

Sepsis Diagnosis and Treatment参考文献 51被引用 39

一句话总结

简要结论: 引入两种方法 MGP-TCN 和 DTW-KNN，用于对不规则时间序列的早期败血症检测；两者均超越此前的最先进方法（MGP-RNN），在发病前最多7小时实现更高的 AUPRC。

ABSTRACT

Sepsis is a life-threatening host response to infection associated with high mortality, morbidity, and health costs. Its management is highly time-sensitive since each hour of delayed treatment increases mortality due to irreversible organ damage. Meanwhile, despite decades of clinical research, robust biomarkers for sepsis are missing. Therefore, detecting sepsis early by utilizing the affluence of high-resolution intensive care records has become a challenging machine learning problem. Recent advances in deep learning and data mining promise to deliver a powerful set of tools to efficiently address this task. This empirical study proposes two novel approaches for the early detection of sepsis: a deep learning model and a lazy learner based on time series distances. Our deep learning model employs a temporal convolutional network that is embedded in a Multi-task Gaussian Process Adapter framework, making it directly applicable to irregularly-spaced time series data. Our lazy learner, by contrast, is an ensemble approach that employs dynamic time warping. We frame the timely detection of sepsis as a supervised time series classification task. For this, we derive the most recent sepsis definition in an hourly resolution to provide the first fully accessible early sepsis detection environment. Seven hours before sepsis onset, our methods improve area under the precision--recall curve from 0.25 to 0.35/0.40 over the state of the art. This demonstrates that they are well-suited for detecting sepsis in the crucial earlier stages when management is most effective.

研究动机与目标

• 解决在不规则采样的 ICU 时间序列中早期败血症检测的挑战。
• 开发端到端模型，考虑采样不规则性，同时最大化预测性能。
• 提供可访问的、按小时的 Sepsis-3 标注框架和面向社区研究的基准数据集。

提出的方法

• 提出 MGP-TCN：将多任务高斯过程适配器与 Temporal Convolutional Networks 相结合，用于不规则采样的多变量时间序列。
• 使用端到端训练来联合优化 MGP 后验和 TCN 分类器。
• 提出 DTW-KNN：基于多变量动态时间规整距离的跨通道集成分类器，通过对每个通道预测取平均来聚合。
• 将评估与 Sepsis-3 的按小时标签及发病前7小时的预测时 horizon 对齐。
• 在 MIMIC-III 数据集上与 MGP-RNN 和 Raw-TCN 基线进行比较，以 AUPRC 为主要指标。
• 使用病例-对照匹配以减少琐碎的判别性并确保现实的比较。

实验结果

研究问题

• RQ1 MGP-TCN 和 DTW-KNN 能否在不规则时间序列上显著早于发病前七小时可靠地检测到败血症？
• RQ2 所提方法是否在按小时 Sepsis-3 框架下优于最先进的 MGP-RNN 和标准 Raw-TCN？
• RQ3 预测时 horizon（最长至发病前7小时）如何影响以 AUPRC 和 AUC 为衡量的模型性能？

主要发现

• MGP-TCN 在发病前七小时将 AUPRC 提升至 0.35，超过先前方法。
• DTW-KNN 集成在发病前七小时将 AUPRC 提升至 0.40，超越当前最佳。
• 两种新方法在各个时 horizon 上均明显优于 MGP-RNN；在此设置下 Raw-TCN 表现不佳。
• 基于插补的 Raw-TCN 性能随着距离发病时间越远而下降，而具不确定性感知的 MGP 基方法保持提升。
• DTW-KNN 在早期表现强劲，但在非常大规模队列上存在可扩展性/内存方面的挑战；MGP-TCN 在在线部署方面更具可扩展性。