arXiv cs.AI· Shahar Oded, Yuval Shahar·· 2 天前AI 评分22
INTERVenE：基于时间抽象区间的 Transformer 用于短期医疗事件预测
INTERVenE: Temporal-Abstraction-Interval Based Transformers for Short-Horizon Medical Event Prediction
AI 导读
INTERVenE 是一族以知识驱动时间抽象（KBTA）为输入的 Transformer 架构，用命名临床概念（状态、趋势、事件、上下文）的 token 流替代原始测量三元组，使逐 token 归因可直接对应临床概念。
正文 · 原文
Abstract:Electronic Health Record (EHR) prediction models in the intensive care unit must learn from sparse and irregular measurements while preserving the clinical meaning of time and supporting transparent decision-making. We present INTERVenE, a family of Transformer architectures whose input is an interval-based, knowledge-based temporal abstraction (KBTA), a token stream of named clinical concepts (states, trends, events, contexts) drawn from a curated medical ontology, rather than an unnamed bin index or a raw measurement triplet. This naming layer is what we ask KBTA to do: it makes the model's per-token attributions resolve to clinical concepts by construction. INTERVenE offers two complementary variants: an auto-regressive decoder that generates future abstraction trajectories with a per-step risk readout (localizing \emph{when} and \emph{after which events} risk rises), and a bidirectional encoder for single-pass joint risk and time-to-event prediction. On 57,078 MIMIC-IV admissions, INTERVenE-Enc outperformed GRU-D, STraTS, and KarmaLego, reaching a support-weighted AUPRC of 0.672 ($+0.046$ over the strongest neural baseline, with non-overlapping 95\% bootstrap CIs), AUROC of 0.901, and length-of-stay MAE of 44.4\,h. INTERVenE-Ar complements these predictions with generated abstraction trajectories and evolving risk scores, achieving AUROC of 0.854 and AUPRC of 0.588 under the harder generative readout. An alternative-discretization ablation preserved predictive performance, supporting KBTA's role as a clinically named attribution substrate rather than a prerequisite for accuracy.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2608.29901 [cs.LG] |
| (or arXiv:2608.29901v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2608.29901 arXiv-issued DOI via DataCite |
Submission history
From: Shahar Oded [view email]
[v1]
Sun, 30 Aug 2026 16:52:34 UTC (861 KB)
[v2]
Thu, 15 Oct 2026 08:52:44 UTC (1,040 KB)
来源：arXiv cs.AI · arxiv.org