Auto-FedRL: Federated Hyperparameter Optimization for Multi-institutional Medical Image Segmentation

被引:14
|
作者
Guo, Pengfei [1 ]
Yang, Dong [2 ]
Hatamizadeh, Ali [2 ]
Xu, An [3 ]
Xu, Ziyue [2 ]
Li, Wenqi [2 ]
Zhao, Can [2 ]
Xu, Daguang [2 ]
Harmon, Stephanie [4 ]
Turkbey, Evrim [5 ]
Turkbey, Baris [4 ]
Wood, Bradford [5 ]
Patella, Francesca [6 ]
Stellato, Elvira [7 ]
Carrafiello, Gianpaolo [7 ]
Patel, Vishal M. [1 ]
Roth, Holger R. [2 ]
机构
[1] Johns Hopkins Univ, Baltimore, MD 21218 USA
[2] NVIDIA, Santa Clara, CA USA
[3] Univ Pittsburgh, Pittsburgh, PA USA
[4] NCI, Bethesda, MD 20892 USA
[5] NIH, Bldg 10, Bethesda, MD 20892 USA
[6] ASST Santi Paolo & Carlo, Milan, Italy
[7] Univ Milan, Milan, Italy
来源
关键词
FL; Reinforcement learning; Hyperparameter optimization;
D O I
10.1007/978-3-031-19803-8_26
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Federated learning (FL) is a distributed machine learning technique that enables collaborative model training while avoiding explicit data sharing. The inherent privacy-preserving property of FL algorithms makes them especially attractive to the medical field. However, in case of heterogeneous client data distributions, standard FL methods are unstable and require intensive hyperparameter tuning to achieve optimal performance. Conventional hyperparameter optimization algorithms are impractical in real-world FL applications as they involve numerous training trials, which are often not affordable with limited compute budgets. In this work, we propose an efficient reinforcement learning (RL)-based federated hyperparameter optimization algorithm, termed Auto-FedRL, in which an online RL agent can dynamically adjust hyperparameters of each client based on the current training progress. Extensive experiments are conducted to investigate different search strategies and RL agents. The effectiveness of the proposed method is validated on a heterogeneous data split of the CIFAR-10 dataset as well as two real-world medical image segmentation datasets for COVID-19 lesion segmentation in chest CT and pancreas segmentation in abdominal CT.
引用
收藏
页码:437 / 455
页数:19
相关论文
共 50 条
  • [1] FedGrav: An Adaptive Federated Aggregation Algorithm for Multi-institutional Medical Image Segmentation
    Deng, Zhifang
    Li, Dandan
    Tan, Shi
    Fu, Ying
    Yuan, Xueguang
    Huang, Xiaohong
    Zhang, Yong
    Zhou, Guangwei
    [J]. MEDICAL IMAGE COMPUTING AND COMPUTER ASSISTED INTERVENTION, MICCAI 2023, PT II, 2023, 14221 : 170 - 180
  • [2] Communication-Efficient Federated Learning for Multi-Institutional Medical Image Classification
    Zhou, Shuang
    Landman, Bennett A.
    Huo, Yuankai
    Gokhale, Aniruddha
    [J]. MEDICAL IMAGING 2022: IMAGING INFORMATICS FOR HEALTHCARE, RESEARCH, AND APPLICATIONS, 2022, 12037
  • [3] Multi-institutional PET/CT image segmentation using federated deep transformer learning
    Shiri, Isaac
    Razeghi, Behrooz
    Sadr, Alireza Vafaei
    Amini, Mehdi
    Salimi, Yazdan
    Ferdowsi, Sohrab
    Boor, Peter
    Guenduez, Deniz
    Voloshynovskiy, Slava
    Zaidi, Habib
    [J]. COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE, 2023, 240
  • [4] Decentralized Distributed Multi-institutional PET Image Segmentation Using a Federated Deep Learning Framework
    Shiri, Isaac
    Sadr, Alireza Vafaei
    Amini, Mehdi
    Salimi, Yazdan
    Sanaat, Amirhossein
    Akhavanallaf, Azadeh
    Razeghi, Behrooz
    Ferdowsi, Sohrab
    Saberi, Abdollah
    Arabi, Hossein
    Becker, Minerva
    Voloshynovskiy, Slava
    Gunduz, Deniz
    Rahmim, Arman
    Zaidi, Habib
    [J]. CLINICAL NUCLEAR MEDICINE, 2022, 47 (07) : 606 - 617
  • [5] Multi-Institutional PET/CT Image Segmentation Using a Decentralized Federated Deep Transformer Learning Algorithm
    Shiri, Isaac
    Amini, Mehdi
    Salimi, Yazdan
    Sanaat, Amirhossein
    Saberi, Abdollah
    Razeghi, Behrooz
    Ferdowsi, Sohrab
    Sadr, Alireza Vafaei
    Voloshynovskiy, Slava
    Gunduz, Deniz
    Rahmim, Arman
    Zaidi, Habib
    [J]. JOURNAL OF NUCLEAR MEDICINE, 2022, 63
  • [6] A multi-institutional meningioma MRI dataset for automated multi-sequence image segmentation
    Labella, Dominic
    Khanna, Omaditya
    McBurney-Lin, Shan
    Mclean, Ryan
    Nedelec, Pierre
    Rashid, Arif S.
    Tahon, Nourel hoda
    Altes, Talissa
    Baid, Ujjwal
    Bhalerao, Radhika
    Dhemesh, Yaseen
    Floyd, Scott
    Godfrey, Devon
    Hilal, Fathi
    Janas, Anastasia
    Kazerooni, Anahita
    Kent, Collin
    Kirkpatrick, John
    Kofler, Florian
    Leu, Kevin
    Maleki, Nazanin
    Menze, Bjoern
    Pajot, Maxence
    Reitman, Zachary J.
    Rudie, Jeffrey D.
    Saluja, Rachit
    Velichko, Yury
    Wang, Chunhao
    Warman, Pranav I.
    Sollmann, Nico
    Diffley, David
    Nandolia, Khanak K.
    Warren, Daniel, I
    Hussain, Ali
    Fehringer, John Pascal
    Bronstein, Yulia
    Deptula, Lisa
    Stein, Evan G.
    Taherzadeh, Mahsa
    Portela de Oliveira, Eduardo
    Haughey, Aoife
    Kontzialis, Marinos
    Saba, Luca
    Turner, Benjamin
    Bruesseler, Melanie M. T.
    Ansari, Shehbaz
    Gkampenis, Athanasios
    Weiss, David Maximilian
    Mansour, Aya
    Shawali, Islam H.
    [J]. SCIENTIFIC DATA, 2024, 11 (01)
  • [7] Federated Cross Learning for Medical Image Segmentation
    Xu, Xuanang
    Deng, Hannah H.
    Chen, Tianyi
    Kuang, Tianshu
    Barber, Joshua C.
    Kim, Daeseung
    Gateno, Jaime
    Xia, James J.
    Yan, Pingkun
    [J]. MEDICAL IMAGING WITH DEEP LEARNING, VOL 227, 2023, 227 : 1441 - 1452
  • [8] Efficient Population Based Hyperparameter Scheduling for Medical Image Segmentation
    He, Yufan
    Yang, Dong
    Myronenko, Andriy
    Xu, Daguang
    [J]. MEDICAL IMAGE COMPUTING AND COMPUTER ASSISTED INTERVENTION, MICCAI 2022, PT V, 2022, 13435 : 560 - 569
  • [9] Collaborative Multi-Institutional Prostate Lesion Segmentation from MR images Using Deep Federated Learning Framework
    Shiri, Isaac
    Showkatian, Eman
    Mohammadi, Reza
    Razeghi, Behrooz
    Bagheri, Soroush
    Hajianfar, Ghasem
    Salimi, Yazdan
    Amini, Mehdi
    Oghli, Mostafa Ghelich
    Ferdowsi, Sohrab
    Voloshynovskiy, Slava
    Zaidi, Habib
    [J]. 2022 IEEE NSS/MIC RTSD - IEEE Nuclear Science Symposium, Medical Imaging Conference and Room Temperature Semiconductor Detector Conference, 2022,
  • [10] Federated Contrastive Learning for Volumetric Medical Image Segmentation
    Wu, Yawen
    Zeng, Dewen
    Wang, Zhepeng
    Shi, Yiyu
    Hu, Jingtong
    [J]. MEDICAL IMAGE COMPUTING AND COMPUTER ASSISTED INTERVENTION - MICCAI 2021, PT III, 2021, 12903 : 367 - 377