Multi-TA: Multilevel Temporal Augmentation for Robust Septic Shock Early Prediction

Authors: Hyunwoo Sohn, Kyungjin Park, Baekkwan Park, Min Chi

IJCAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Validated on two real-world EHRs for septic shock, Multi-TA outperforms mixup and GAN-based stateof-the-art models across eight prediction windows, demonstrating its temporal robustness.
Researcher Affiliation Collaboration Hyunwoo Sohn1 , Kyungjin Park2 , Baekkwan Park3 and Min Chi1 1North Carolina State University 2Integral Ad Science 3University of Missouri
Pseudocode Yes Algorithm 1: Trajectory-level Augmentation
Open Source Code No The paper does not provide an explicit statement or link indicating that the source code for the methodology is openly available.
Open Datasets No We utilize two real-world EHR datasets collected from 210,289 and 106,844 adult patient visits to Christiana Care Health System (CCHS) and Mayo Clinic (Mayo), respectively. Both datasets comprise 2.5 years (07/2013-12/2015) of anonymized and institutional review board (IRB)-approved EHRs.
Dataset Splits Yes Regular is a commonly used test set for early prediction, in which the numbers of patients in training, validation, and test sets decrease as prediction hour n increases.
Hardware Specification No The paper does not provide specific details about the hardware (e.g., GPU/CPU models, memory) used for running the experiments.
Software Dependencies No The paper mentions software components like BERT and LSTM but does not provide specific version numbers for any software libraries or dependencies used in the experiments.
Experiment Setup No The paper mentions that 'Hyperparameters, such as p and U, are chosen based on the validation performance' and lists 'learning rate η' as a parameter in Algorithm 1, but it does not provide the concrete numerical values for these or other training parameters like batch size, number of epochs, or optimizer settings.