Frontiers

A multi-graph convolutional network method for Alzheimer’s disease diagnosis based on multi-frequency EEG data with dual-mode connectivity

Objective: Alzheimer’s disease (AD) is mainly identified by cognitive function deterioration. Diagnosing AD at early stages poses significant challenges for both researchers and healthcare ...

Interpretable neural networks help reveal the nature of dark matter

A research team from the Xinjiang Astronomical Observatory (XAO) of the Chinese Academy of Sciences has developed an ...
IEEE

A Nonlinear Weighted Graph Convolution Network Based on Manifold Geometric Regularization for Hyperspectral Image Classification

Abstract: Extracting spatial–spectral joint features has become a critical approach for improving model classification performance in the field of hyperspectral image classification (HSIC). However, ...
Search Engine Land

Google’s great clarity cleanup: 3 shifts redefining the Knowledge Graph and its AI future

Google’s Knowledge Graph saw its largest contraction in a decade in June: a two-stage, one-week drop of 6.26% – over 3 billion entities deleted. Since 2015, we’ve tracked the Knowledge Graph and have ...
IEEE

Augmented Graph Convolutional Network for Enhancing Label Reachability

Abstract: Graph Convolutional Networks (GCNs) have emerged as a leading approach for semi-supervised node classification. However, due to the uneven distribution of labeled nodes in graphs, only a ...