Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: May 2, 2026

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

996

SS-AdaMoE: Spatio-Spectral Adaptive Mixture of Experts with Global Structural Priors for Graph Node Classification.

Xilin Kang1, Tianyue Yu1, Letao Wang2

  • 1School of Computer, Jiangsu University of Science and Technology, Zhenjiang 212100, China.

Entropy (Basel, Switzerland)
|March 28, 2026
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

Structural Classification of Joints01:20

Structural Classification of Joints

8.7K
Joints, also known as articulations, are classified based on their structural characteristics, i.e., based on whether the articulating surfaces of the adjacent bones are directly connected by fibrous connective tissue or cartilage, or whether the articulating surfaces contact each other within a fluid-filled joint cavity. These differences serve to divide the joints of the body into three structural classifications.
A fibrous joint is where the adjacent bones are united by fibrous connective...
8.7K

You might also read

Related Articles

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

Sort by
Same author

Correction: Zhang et al. Caudate-Centric Triphasic Network Reconfiguration Characterizes the Early Progression of Cognitive Impairment in Parkinson's Disease: A Simultaneous PET/fMRI Study. Journal of Integrative Neuroscience. 2026; 25(2): 46634.

Journal of integrative neuroscience·2026
Same author

Career Adaptability and Academic Achievement Among Chinese High School Students: A Three-Wave Longitudinal Study of Social Cognitive and Metacognitive Mediating Mechanisms.

Journal of Intelligence·2026
Same author

SlABH15 Promotes Jasmonate-Mediated Lead Detoxification in Tomato: A Promising Target for Enhancing Phytoremediation and Lead Stress Tolerance.

Journal of agricultural and food chemistry·2026
Same author

DRP1/p53 interaction-mediated mitochondrial regulation of TFRC-SLC39A14 drives mechanical overload-induced temporomandibular joint osteoarthritis via ferroptosis.

Clinical science (London, England : 1979)·2026
Same author

The Upregulation of AIM2 in the Central Nucleus of the Amygdala Correlates with Pain Induced by Tooth Movement.

International journal of molecular sciences·2026
Same author

Bionic design of gas sensors for selective switching and humidity suppression.

Biosensors & bioelectronics·2026

This study introduces SS-AdaMoE, a new Graph Neural Network (GNN) framework that improves node classification on complex graphs. It effectively captures high-frequency signals and uses global context for better performance.

Area of Science:

  • Graph Neural Networks
  • Machine Learning
  • Network Science

Background:

  • Traditional Graph Neural Networks (GNNs) act as low-pass filters, limiting their effectiveness on heterophilic graphs where high-frequency signals are important.
  • Existing Mixture-of-Experts (MoE) methods for graphs often rely on local information, failing to incorporate global structural context for expert selection.

Purpose of the Study:

  • To propose SS-AdaMoE, a novel Spatio-Spectral Adaptive Mixture of Experts framework for robust node classification on diverse graph structures.
  • To enable adaptive capture of arbitrary frequency responses, including crucial high-pass and band-pass signals, which are typically missed by standard GNNs.
  • To address the locality bias in MoE by incorporating global topological awareness into expert selection.

Main Methods:

Keywords:
Bernstein polynomialsadaptive routinggraph neural networksgraph transformerheterophilic graphsmixture of expertsnode classificationspectral filtering

Related Experiment Videos

Last Updated: May 2, 2026

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

996
  • Developed a Dual-Domain Expert System integrating spatial aggregators with learnable spectral filters (Bernstein polynomials) for adaptive frequency response.
  • Introduced a Hierarchical Global-Prior Gating Network augmented by a Linear Graph Transformer to guide expert selection using both local features and global topology.
  • Conducted extensive experiments on five benchmark datasets covering homophilic and heterophilic networks.

    • Main Results:

    • SS-AdaMoE demonstrated consistent outperformance across all tested datasets.
    • Achieved accuracy improvements of up to 2.65% on Chameleon and 1.41% on Roman-empire compared to the strongest MoE baseline.
    • Surpassed traditional GCN architectures by over 28% on heterophilic datasets like Texas.

    Conclusions:

    • The synergy of learnable spectral priors and global gating in SS-AdaMoE effectively bridges the gap between spatial aggregation and spectral filtering.
    • SS-AdaMoE offers a robust solution for node classification, particularly excelling in heterophilic graph settings.
    • The proposed framework enhances GNN capabilities by adaptively capturing essential high-frequency signals and leveraging global graph structure.