首页|Eindhoven University of Technology Reports Findings in Machine Learning (An adve rsarial learning approach to generate pressure support ventilation waveforms for asynchrony detection)
Eindhoven University of Technology Reports Findings in Machine Learning (An adve rsarial learning approach to generate pressure support ventilation waveforms for asynchrony detection)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning DailyNews Daily News - New research on Machine Learning is th e subject of a report. According to newsreporting from Eindhoven, Netherlands, by NewsRx journalists, research stated, “Mechanical ventilation isa life-saving treatment for critically-ill patients. During treatment, patient-ventilator asy nchrony (PVA)can occur, which can lead to pulmonary damage, complications, and higher mortality.”The news correspondents obtained a quote from the research from the Eindhoven Un iversity of Technology,“While traditional detection methods for PVAs rely on vi sual inspection by clinicians, in recentyears, machine learning models are bein g developed to detect PVAs automatically. However, trainingthese models require s large labeled datasets, which are difficult to obtain, as labeling is a labour -intensiveand time-consuming task, requiring clinical expertise. Simulating the lung-ventilator interactions has beenproposed to obtain large labeled datasets to train machine learning classifiers. However, the obtained datalacks the inf luence of different hardware, of servo-controlled algorithms, and different sour ces of noise.Here, we propose VentGAN, an adversarial learning approach to impr ove simulated data by learning theventilator fingerprints from unlabeled clinic al data. In VentGAN, the loss functions are designed to addcharacteristics of c linical waveforms to the generated results, while preserving the labels of the s imulatedwaveforms. To validate VentGAN, we compare the performance for detectio n and classification of PVAswhen training a previously developed machine learni ng algorithm with the original simulated data and withthe data generated by Ven tGAN. Testing is performed on independent clinical data labeled by experts.The McNemar test is applied to evaluate statistical differences in the obtained clas sification accuracy.VentGAN significantly improves the classification accuracy for late cycling, early cycling and normal breaths(p <0.01); no significant difference in accuracy was observed for delayed inspirations (p = 0.2), while theaccuracy decreased for ineffective efforts (p <0.01).”
NETL
NSTL
