Dynamic Bank Learning for Semi-supervised Federated Image Diagnosis with Class Imbalance

Jiang, Meirui; Yang, Hongzheng; Li, Xiaoxiao; Liu, Quande; Heng, Pheng‐Ann; Dou, Qi

doi:10.1007/978-3-031-16437-8_19

Cited by 20 publications

(4 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To counter this, the authors decomposed the model parameters θ to two variables θ = ψ + ρ and utilize a separate updating strategy, where only ψ is updated during unsupervised learning, and similarly, ρ is updated for supervised learning. In a real-world scenario, Jiang et al [76] addressed the challenge of imbalanced class distributions among unlabeled clients in the context of medical image diagnosis. They proposed a novel scheme called dynamic bank learning, which aims to collect confident samples and subsequently divide them into sub-banks with varying class proportions.…”

Section: Federated Semi-supervised Learningmentioning

confidence: 99%

Emerging Trends in Federated Learning: From Model Fusion to Federated X Learning

Ji,

Tan,

Saravirta

et al. 2023

Preprint

View full text Add to dashboard Cite

Federated learning is a new learning paradigm that decouples data collection and model training via multi-party computation and model aggregation.As a flexible learning setting, federated learning has the potential to integrate with other learning frameworks.We conduct a focused survey of federated learning in conjunction with other learning algorithms. Specifically, we explore various learning algorithms to improve the vanilla federated averaging algorithm and review model fusion methods such as adaptive aggregation, regularization, clustered methods, and Bayesian methods. Following the emerging trends, we also discuss federated learning in the intersection with other learning paradigms, termed federated X learning, where X includes multitask learning, meta-learning, transfer learning, unsupervised learning, and reinforcement learning. This survey reviews the state of the art, challenges, and future directions.

show abstract

Section: Federated Semi-supervised Learningmentioning

confidence: 99%

Emerging Trends in Federated Learning: From Model Fusion to Federated X Learning

Ji,

Tan,

Saravirta

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…CReRF [33] recalibrates the federated classifier in the server using synthesized balanced features. FedIRM [63] and imFedSemi [62] resolves class imbalance in semi-supervised settings by sharing class relation matrices and highly confident unlabeled samples respectively. However, sharing additional information beyond model weight update is not desirable in medical domains due to the potential risk of privacy leakage [71].…”

Section: B Class Imbalance Learningmentioning

confidence: 99%

Customized Federated Learning for Multi-Source Decentralized Medical Image Classification

Wicaksana

Yan

Yang

et al. 2022

IEEE J. Biomed. Health Inform.

View full text Add to dashboard Cite

The performance of deep networks for medical image analysis is often constrained by limited medical data, which is privacy-sensitive. Federated learning (FL) alleviates the constraint by allowing different institutions to collaboratively train a federated model without sharing data. However, the federated model is often suboptimal with respect to the characteristics of each client's local data. Instead of training a single global model, we propose Customized FL (CusFL), for which each client iteratively trains a client-specific/private model based on a federated global model aggregated from all private models trained in the immediate previous iteration. Two overarching strategies employed by CusFL lead to its superior performance: 1) the federated model is mainly for feature alignment and thus only consists of feature extraction layers; 2) the federated feature extractor is used to guide the training of each private model. In that way, CusFL allows each client to selectively learn useful knowledge from the federated model to improve its personalized model. We evaluated CusFL on multi-source medical image datasets for the identification of clinically significant prostate cancer and the classification of skin lesions.

show abstract

“…A smooth optimum is one in which a neighbourhood region surrounding the local optimal solution has a loss that is consistent over the entire neighbourhood. In addition, the authors of [21] proposed an FL approach that can perform inside and outside model personalization via an insubstantialgradient method to make use of the local customized model by gathering global updates for general expertise and local updates for client-given optimization. Moreover, the authors of [22] identified and solved a novel issue setting known as federated domain generalization to enable learning a federated model from several remote source domains in such a way that it can directly generalise to domains that have not been encountered before.…”

Section: B Federated Learning For Domain Adaptionmentioning

confidence: 99%

Collaborative Screening of COVID-19-like Disease from Multi-Institutional Radiographs: A Federated Learning Approach

2022

View full text Add to dashboard Cite

COVID-19-like pandemics are a major threat to the global health system have the potential to cause high mortality across age groups. The advance of the Internet of Medical Things (IoMT) technologies paves the way toward developing reliable solutions to combat these pandemics. Medical images (i.e., X-rays, computed tomography (CT)) provide an efficient tool for disease detection and diagnosis. The cost, time, and efforts for acquiring and annotating, for instance, large CT datasets make it complicated to obtain large numbers of samples from a single institution. However, owing to the necessity to preserve the privacy of patient data, it is challenging to build a centralized dataset from many institutions, especially during a pandemic. Moreover, heterogeneity between institutions presents a barrier to building efficient screening solutions. Thus, this paper presents a fog-based federated generative domain adaption framework (FGDA), where fog nodes aggregate patients’ data necessary to collaboratively train local deep-learning models for disease screening in medical images from different institutions. Local differential privacy is presented to protect the local gradients against attackers during the global model aggregation. In FGDA, the generative domain adaptation (DA) method is introduced to handle data discrepancies. Experimental evaluation on a case study of COVID-19 segmentation demonstrated the efficiency of FGDA over competing learning approaches with statistical significance.

show abstract

Dynamic Bank Learning for Semi-supervised Federated Image Diagnosis with Class Imbalance

Cited by 20 publications

References 16 publications

Emerging Trends in Federated Learning: From Model Fusion to Federated X Learning

Emerging Trends in Federated Learning: From Model Fusion to Federated X Learning

Customized Federated Learning for Multi-Source Decentralized Medical Image Classification

Collaborative Screening of COVID-19-like Disease from Multi-Institutional Radiographs: A Federated Learning Approach

Contact Info

Product

Resources

About