计算机科学

计算机科学 ›› 2023, Vol. 50 ›› Issue (10): 165-175.doi: 10.11896/jsjkx.220900177

• 人工智能 • 上一篇    下一篇

基于联邦学习的网络化ICU呼吸机和镇静剂管理方法

曹林霄1, 刘佳2, 朱怡飞3, 周浩泉4, 龚伟4, 于卫华5, 李朝友6   

  1. 1 中国科学技术大学大数据学院 合肥230026
    2 安徽医科大学附属省立医院儿科 合肥230001
    3 上海交通大学密西根学院 上海200240
    4 中国科学技术大学附属第一医院儿科 合肥230001
    5 合肥市第一人民医院护理部 合肥230061
    6 合肥市第一人民医院儿科 合肥230061
  • 收稿日期:2022-09-17 修回日期:2023-02-10 出版日期:2023-10-10 发布日期:2023-10-10
  • 通讯作者: 周浩泉(zhouhq2005@qq.com)
  • 作者简介:(linxiaocao@mail.ustc.edu.cn)
  • 基金资助:
    合肥市自主创新政策“借转补”项目(J2020Y03)

Ventilator and Sedative Management in Networked ICUs Based on Federated Learning

CAO Linxiao1, LIU Jia2, ZHU Yifei3, ZHOU Haoquan4, GONG Wei4, YU Weihua5, LI Chaoyou6   

  1. 1 School of Data Science,University of Science and Technology of China,Hefei 230026,China
    2 Department of Pediatrics,Provincial Hospital Affiliated to Anhui Medical University,Hefei 230001,China
    3 University of Michigan-Shanghai Jiao Tong University Joint Institute,Shanghai Jiao Tong University,Shanghai 200240,China
    4 Department of Pediatrics,The First Affiliated Hospital of USTC,Hefei 230001,China
    5 Department of Nursing,The First People's Hospital of Hefei,Hefei 230061,China
    6 Department of Pediatrics,The First People's Hospital of Hefei,Hefei 230061,China
  • Received:2022-09-17 Revised:2023-02-10 Online:2023-10-10 Published:2023-10-10
  • About author:CAO Linxiao,born in 2000,postgra-duate.His main research interests include federated learning and reinforcement learning.ZHOU Haoquan,born in 1968,master,chief physician.His main research interests include childhood respiratory di-sease,bronchoscopy,and smart healthcare.
  • Supported by:
    Hefei Healthcare Grant(J2020Y03).

PDF (PC)

1971

摘要： 医疗物联网设备的激增和丰富的医疗数据为智慧医疗提供了新的可能。重症监护室(ICU)的病人依靠众多医疗边缘设备来持续监测管理患者的健康状况。在ICU常见的治疗干预措施中,有创机械通气和镇静剂的注射多用于维持患者的呼吸功能,提高治疗质量,而现有的治疗干预措施很大程度上依赖于医生的判断。文中提出了一种基于联邦学习的临床辅助决策方法——MFed,可以基于网络化ICU分布式协作学习最佳干预政策。该方法应用基于差分隐私的联邦学习方法,打破了医疗数据隐私方面的限制以及医疗数据孤岛的窘境;用分布鲁棒优化确保最坏情况下的性能并结合伪孪生网络实现自适应地滤除噪声数据。最后,在现实ICU数据集上的实验表明,与其他最先进的基线相比,所提方法的准确率提高了36.75％。

关键词: 联邦学习, 医疗物联网, 分布式鲁棒优化, 医疗数据噪声, 医疗数据隐私

Abstract: The proliferation of medical IoT devices and the abundance of medical data open up new possibilities for smart healthcare.Patients in the intensive care unit (ICU) rely on numerous medical IoT devices to continuously monitor and manage patients' health status.Among the common therapeutic interventions in ICUs,invasive mechanical ventilation and sedation are mostly administered to maintain patients' respiratory function and enhance the care quality.While the existing therapeutic interventions rely heavily on physician judgment.This paper proposes a data-driven optimal policy learning framework named MFed that allows distributed learning of optimal intervention policies on networked ICUs.A differentially private federated learning method is constructed to overcome privacy limitations in medical data.MFed further ensures worst-case performance with distributionally robust optimization and adaptively filters out noisy data.Extensive experiments on a real-world ICU dataset show that the proposed method improves accuracy by 36.75% compared to other state-of-the-art baselines.

Key words: Federated learning, Medical IoT, Distributionally robust optimization, Medical data noise, Medical data privacy

中图分类号: 

  • TP391
[1]DELL TECHNOLOGIES,Dell emc global data protection index survey[EB/OL].[2022-08-10].https://www.dellemc.com/en-us/data-protection/gdpi/index.htm.
[2]HUESCH M D,MOSHER T J.Using it or losing it?the casefor data scientists inside health care[J／OL].Nejm Catalyst,2017,3(3).https://catalyst.nejm.org/doi/full/10.1056/CAT.17.0493.
[3]RAZA M,AWAIS M,SINGH N,et al.Intelligent iot frameworkfor indoor healthcare monitoring of parkinson's disease patient[J].IEEE Journal on Selected Areas in Communications,2020,39(2):593-602.
[4]CONTI G,MANTZ J,LONGROIS D,et al.Sedation and weaning from mechanical ventilation:time for ‘best practice'to catch up with new realities?[J].Multidisciplinary Respiratory Medicine,2014,9(1):1-5.
[5]AMBROSINO N,GABBRIELLI L.The difficult-to-wean patien[J].Expert Review of Respiratory Medicine,2010,4(5):685-692
[6]PATEL S B,KRESS J P.Sedation and analgesia in the mechanically ventilated patient[J].American Journal of Respiratory and Critical Care Medicine,2012,185(5):486-497.
[7]ZHANG Y,HE D,OBAIDAT M S,et al.Efficient identity-based distributed decryption scheme for electronic personal health record sharing system[J].IEEE Journal on Selected Areas in Communications,2020,39(2):384-395.
[8]WANG J Z,KONG L W,HUANG Z C,et al.Research progress on privacy protection in federated learning [J].Big Data,2021,7(3):130-149.
[9]MCMAHAN B,MOORE E,RAMAGE D,et al.Communica-tion-efficient learning of deep networks from decentralized data[C]//Artificial Intelligence and Statistics.PMLR,2017:1273-1282.
[10]ZHANG D Y,KOU Z,WANG D.Fedsens:A federated learning approach for smart health sensing with class imbalance in resource constrained edgecomputing[C]//IEEE INFOCOM 2021－IEEE Conference on Computer Communications.IEEE,2021:1-10.
[11]ELAYAN H,ALOQAILY M,GUIZANI M.Sustainability ofhealthcare data analysis IoT-based systems using deep federated learning[J].IEEE Internet of Things Journal,2021,9(10):7338-7346.
[12]YAN Z,WICAKSANA J,WANG Z,et al.Variation-aware fe-derated learning with multi-source decentralized medical image data[J].IEEE Journal of Biomedical and Health Informatics,2020,25(7):2615-2628.
[13]SHELLER M J,REINA G A,EDWARDS B,et al.Multiinstitutional deep learning modeling w/o sharing patient data:A feasibility study on brain tumor segmentation[J].MICCAI Brainlesion Workshop,2019:92-104.
[14]ZHANG W,ZHOU T,LU Q,et al.Dynamic-fusion-based fede-rated learning for covid-19 detection[J].IEEE Internet of Things Journal,2021,8(21):15884-15891.
[15]GUO P,WANG P,ZHOU J,et al.Multi-institutional collaborations for improving deep learning-based magnetic resonance image reconstruction using federated learning[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2021:2423-2432.
[16]MALEKZADEH M,HASIRCIOGLU B,MITAL N,et al.Dopa-mine:Differentially private federated learning on medical data[J].arXiv:2101.11693,2021.
[17]LU S,ZHANG Y,WANG Y,et al.Learn electronic health records by fully decentralized federated learning[J].arXiv:1912.01792,2019.
[18]WU Q,CHEN X,ZHOU Z,et al.Fedhome:Cloud-edge basedpersonalized federated learning for in-home health monitoring[J].IEEE Transactions on Mobile Computing,2020,21(8):2818-2832.
[19]XU X,PENG H,SUN L,et al.Fedmood:Federated learning on mobile health data for mood detection[J].arXiv:2102.09342,2021.
[20]TAN X,CHANG C C H,TANG L.A tree-based federatedlearning approach for personalized treatment effect estimation from heterogeneous data sources[C]//International Conference on Machine Learning.PMLR,2022:21013-21036.
[21]ELAYAN H,ALOQAILY M,GUIZANI M.Deep federatedlearning for iot-based decentralized healthcare systems[C]//2021 International Wireless Communications and Mobile Computing(IWCMC).IEEE,2021:105-109.
[22]LIU Y X,CHEN H,LIU Y H,et al.Privacy-Preserving Technologies in Federated Learning [J].Journal of Software,2022,33(3):1057-1092.
[23]YANG G,WANG Z S.Research progress of privacy protection in federated learning[J].Journal of Nanjing University of Posts and Telecommunications(Natural Science Edition),2020,40(5):204-214.
[24]INAN A,GURSOY M E,SAYGIN Y.Sensitivity analysis fornon-interactive differential privacy:bounds and efficient algorithms[J].IEEE Transactions on Dependable and Secure Computing,2017,17(1):194-207.
[25]CHOUDHURY O,GKOULALAS-DIVANIS A,SALONIDIST,et al.Differential privacy-enabled federated learning for sensitive health data[J].arXiv:1910.02578,2019.
[26]ADNAN M,KALRA S,CRESSWELL J C,et al.Federatedlearning and differential privacy for medical image analysis[J].Scientific Reports,2022,12(1):1-10.
[27]LEE C,LUO Z,NGIAM K Y,et al.Big healthcare data analy-tics:Challenges and applications[M]//Handbook of Large-scale Distributed Computing in Smart Healthcare.2017:11-41.
[28]HAN B,YAO Q,YU X,et al.Co-teaching:Robust training ofdeep neural networks with extremely noisy labels[J].Advances in Neural Information Processing Systems,2018,31:8535-8545.
[29]WEI H,FENG L,CHEN X,et al.Combating noisy labels byagreement:A joint training method with co-regularization [C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2020:13726-13735.
[30]PRASAD N,CHENG L F,CHIVERS C,et al.A reinforcement learning approach to weaning of mechanical ventilation in intensive care units[J].arXiv:1704.06300,2017.
[31]YU C,LIU J,ZHAO H.Inverse reinforcement learning for intelligent mechanical ventilation and sedative dosing in intensive care units[J].BMC Medical Informatics and Decision Making,2019,19(2):111-120.
[32]SUTTON R S,BARTO A G.Reinforcement learning:An introduction[M].MIT Press,2018.
[33]NG A Y,RUSSELL S.Algorithms for inverse reinforcementlearning[C]//ICML.2000,1:2.
[34]CHAN J,VAN DER SCHAAR M.Scalable bayesian inverse reinforcement learning[J].arXiv:2102.06483,2021.
[35]DENG Y,KAMANI M M,MAHDAVI M.Distributionally robust federated averaging[J].Advances in Neural Information Processing Systems,2020,33:15111-15122.
[36]JOHNSON E,POLLARD T J,SHEN L,et al.Mimic-iii,a freely accessible critical care database[J].Scientific Data,2016,3(1):1-9.
[37]ERNST D,STAN G B,GONCALVES J,et al.Clinical databased optimal sti strategies for hiv:a reinforcement learning approach[C]//Proceedings of the 45th IEEE Conference on Decision and Control.IEEE,2006:667-672.
[38]RIEDMILLER M.Neural fitted q iteration-first experienceswith a data efficient neural reinforcement learning method [C]//European Conference on Machine Learning.2005:317-328.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发