Ehrs
PulseAugur coverage of Ehrs — every cluster mentioning Ehrs across labs, papers, and developer communities, ranked by signal.
1 天有情绪数据
-
New MedTPE method compresses EHR data for LLMs with no performance loss
Researchers have developed a new method called Medical Token-Pair Encoding (MedTPE) to efficiently compress long electronic health record sequences for large language models. This technique merges frequently occurring m…
-
AI models predict patient risk using clinical notes and temporal data
Researchers have developed two novel methods, HiTGNN and ReVeAL, to improve early risk prediction for chronic diseases using clinical language processing. HiTGNN, a hierarchical temporal graph neural network, effectivel…
-
LLM 和归一化流用于处理不完整的医疗保健数据以估计治疗效果
研究人员开发了一种新颖的两阶段流程 CausalFlow-T，旨在从不完整的纵向电子健康记录中改进治疗效果估计。第一阶段利用具有 LSTM 编码的 DAG 约束归一化流进行精确的反事实推断，第二阶段则采用 LLM 驱动的插补器来处理缺失数据。与统计基线相比，这种组合方法在各种缺失水平下均能更优地保留平均治疗效果的恢复。
-
Study: Shorter data windows optimize AI for hospital readmission prediction
A new study published on arXiv explores the optimal historical data window for predicting hospital readmissions. Researchers found that for unstructured clinical notes, a shorter window of three to six months prior to s…
-
稀疏自编码器分解临床序列模型表征：特征复杂度、任务专业化与死亡率预测
研究人员开发了几种利用电子健康记录（EHR）上的机器学习来改进临床预测的新方法。其中一种方法“Risk Horizons”使用一种具有几何感知的框架和双曲嵌入来构建患者特定的候选空间，以预测未来的临床事件。另一种方法将临床诊断视为一个自回归序列建模任务，采用大型语言模型的因果解码器来处理缺失的模态并提高可解释性。此外，一个名为FlatASCEND的新模型专注于具有连续时间预测的自回归临床序列生成，并测试药理学关联，而另一项研究则使用稀…