Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Videos

FedSKD: Aggregation-Free Model-Heterogeneous Federated Learning via Multidimensional Similarity Knowledge

Ziqiao Weng, Weidong Cai, Bo Zhou

IEEE Transactions on Neural Networks and Learning Systems

|April 22, 2026

Summary

This summary is machine-generated.

Related Concept Videos

Aggregates Classification

Aggregates Classification

Aggregate classification is generally based on its size, petrographic characteristics, weight, and source. Size classification ranges from coarse to fine aggregates, defined by the size of the particles. Coarse aggregates are particles that do not pass through ASTM sieve No. 4, and aggregates that pass through the sieve are fine aggregates.
Petrographic classification groups aggregates based on common mineralogical characteristics. Some of the common mineral groups found in aggregates are...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Regulation of microtubule dynamics by DIAPH3 influences amoeboid tumor cell mechanics and sensitivity to taxanes.

Scientific reports·2015

Same author

Aberrant Functional Connectivity Architecture in Alzheimer's Disease and Mild Cognitive Impairment: A Whole-Brain, Data-Driven Analysis.

BioMed research international·2015

Same author

Alternative NF-κB Isoforms in the Drosophila Neuromuscular Junction and Brain.

PloS one·2015

Same author

Grape seed proanthocyanidin protects liver against ischemia/reperfusion injury by attenuating endoplasmic reticulum stress.

World journal of gastroenterology·2015

Same author

Serum Levels of Progranulin Are Closely Associated with Microvascular Complication in Type 2 Diabetes.

Disease markers·2015

Same author

Repression of microRNA biogenesis by silencing of OsDCL1 activates the basal resistance to Magnaporthe oryzae in rice.

Plant science : an international journal of experimental plant biology·2015

Same journal

Intervention Feasible Region and Driver Risk Capacity Aware Human-Machine Collaborative Safe Trajectory Planning.

IEEE transactions on neural networks and learning systems·2026

Same journal

A Unified Differential Denoising Learning Framework With a Pre-Trained Model and Fuzzy Graph Networks for Drug-Drug Interaction Prediction.

IEEE transactions on neural networks and learning systems·2026

Same journal

Self-Supervised Continuous Dynamic Graph Representation Learning via Hawkes Processes.

IEEE transactions on neural networks and learning systems·2026

Same journal

cPU: Consistent Risk Estimator for Positive-Unlabeled Learning.

IEEE transactions on neural networks and learning systems·2026

Same journal

Tuning-Free Latent Diffusion Models for Ultrahigh-Resolution Image Editing.

IEEE transactions on neural networks and learning systems·2026

Same journal

Hidden Data Recovery and Forecasting via Next-Generation Reservoir Computing With Multiscale Delay Selection.

IEEE transactions on neural networks and learning systems·2026

See all related articles

Federated learning (FL) with heterogeneous models is enhanced by FedSKD, a peer-to-peer framework enabling direct knowledge sharing. This approach improves medical image classification accuracy and generalization without central servers.

Area of Science:

Artificial Intelligence
Machine Learning
Medical Imaging

Background:

Federated learning (FL) allows privacy-preserving model training without data sharing.
Model-heterogeneous FL (MHFL) permits diverse client model architectures but faces scalability issues with centralized aggregation or similar architectures.
Existing peer-to-peer (P2P) FL lacks support for heterogeneous models and suffers from model drift and knowledge dilution.

Purpose of the Study:

To introduce FedSKD, a novel P2P MHFL framework for medical image classification.
To enable direct, serverless knowledge exchange among clients with fully heterogeneous model architectures.
To address challenges of scalability, efficiency, model drift, and knowledge dilution in P2P MHFL.

Main Methods:

FedSKD utilizes a round-robin model circulation for direct knowledge exchange.

Related Experiment Videos

The core innovation is multidimensional similarity knowledge distillation (SKD) for bidirectional, cross-client knowledge transfer.

Knowledge transfer occurs at batch, pixel/voxel, and region levels, mitigating forgetting and drift while preserving heterogeneity.

Main Results:

FedSKD demonstrated superior performance compared to state-of-the-art heterogeneous and homogeneous FL baselines.
Evaluations on fMRI-based autism spectrum disorder diagnosis and skin lesion classification showed improved personalization and generalization.
The framework effectively mitigates catastrophic forgetting and model drift through progressive reinforcement and distribution alignment.

Conclusions:

FedSKD offers a scalable and robust solution for P2P MHFL in medical imaging.
The multidimensional SKD approach successfully facilitates knowledge transfer between heterogeneous models.
This framework enhances cross-institutional generalization and model personalization in decentralized learning settings.