Graph neural networks (GNNs) have achieved great success in many graph-based applications. However, the enormous size and high sparsity level of graphs hinder their applications under industrial scenarios. Although some scalable GNNs are proposed for large-scale graphs, they adopt a fixed $K$-hop neighborhood each node, thus facing over-smoothing issue when adopting large propagation depths nodes within sparse regions. To tackle above issue, we propose new GNN architecture -- Attention...

Graph Neural Networks (GNNs) have achieved great success in various graph mining tasks.However, drastic performance degradation is always observed when a GNN stacked with many layers. As result, most GNNs only shallow architectures, which limits their expressive power and exploitation of deep neighborhoods.Most recent studies attribute the to \textit{over-smoothing} issue. In this paper, we disentangle conventional convolution operation into two independent operations: \textit{Propagation}...

Abstract Class imbalance is a prevalent issue in practical applications, which poses significant challenges for classifiers. The large margin distribution machine (LDM) introduces the of samples to replace traditional minimum margin, resulting extensively enhanced classification performance. However, hyperplane LDM tends be skewed toward minority class, due optimization property means. Moreover, absence non-deterministic options and measurement confidence level further restricts capability...

Graph neural networks (GNNs) have been widely adopted for semi-supervised learning on graphs. A recent study shows that the graph random network (GRAND) model can generate state-of-the-art performance this problem. However, it is difficult GRAND to handle large-scale graphs since its effectiveness relies computationally expensive data augmentation procedures. In work, we present a scalable and high-performance GNN framework GRAND+ learning. To address above issue, develop generalized forward...

The constant need for decarbonization has led to the replacement of artificial light at night (ALAN) with light-emitting diodes (LEDs), inducing blue pollution and its consequent adverse effects. As a result, there is an urgent development technique rapid, accurate, large-scale discrimination various illumination sources. newly launched Sustainable Development Science Satellite-1 (SDGSAT-1) can play this role by supplementing existing nighttime data multispectral high-resolution features....

Purpose With the rapid development of economy, carbon emissions have also risen sharply. This study explores relationship between two by combining literature relevant fields and maps analytical framework from knowledge base to research frontier model using CiteSpace. Design/methodology/approach Using CiteSpace data statistical tools, we conducted a bibliometric visual analysis nearly ten thousand papers on economic published in Web Science (WOS) China National Knowledge Infrastructure (CNKI)...

Maximal biclique enumeration is a fundamental problem in bipartite graph data analysis. Existing methods mainly focus on non-attributed graphs and also ignore the fairness of attributes. In this paper, we introduce concept into model for first time study fairness-aware enumeration. Specifically, propose two models, called single-side fair bi-side respectively. To efficiently enumerate all bicliques, present non-trivial pruning techniques, α-β core colorful pruning, to reduce size without...

In this paper, an extreme learning machine (ELM) network based on improved shuffled frog leaping algorithm (CCSFLA) is applied in early bearing fault diagnosis. ELM a new type of single layer forward network. Although the generalization stronger compared with traditional neural networks, random setup initial parameters increases instability An SFLA sinusoidal chaotic mapping infinite collapses and constriction factors proposed paper to optimize obtain CCSFLA–ELM model. Results show that...

In recent years, through the implementation of a series policies, such as delimitation major grain producing areas and construction advantageous characteristic agricultural product areas, spatial distribution agriculture in China has changed significantly; however, research on impact changes efficiency technology is still lacking. Taking 11 cities Hebei Province object, this study examines dependence regional technical using stochastic frontier analysis econometric analysis. The results show...

In this paper, an extreme learning machine (ELM) network based on improved shuffled frog leaping algorithm (CCSFLA) is applied in early bearing fault diagnosis. ELM a new type of single layer...

Graph neural networks (GNNs) have recently achieved state-of-the-art performance in many graph-based applications. Despite the high expressive power, they typically need to perform an expensive recursive neighborhood expansion multiple training epochs and face a scalability issue. Moreover, most of them are inflexible since restricted fixed-hop neighborhoods insensitive actual receptive field demands for different nodes. We circumvent these limitations by introducing scalable flexible...

Young rural women who migrate to China’s large cities for work (Dagongmei) occupy a liminal position in space and time that is conditioned by particularly gendered form of mobility relationship place(s). As migrants, they live away from their homes but are not fully incorporated into urban society, anticipate eventual return. unmarried women, poised between childhood adulthood. Their unique creates vulnerability gives rise feelings ambivalence. This article reading Collins’s concept...

Maximal biclique enumeration is a fundamental problem in bipartite graph data analysis. Existing methods mainly focus on non-attributed graphs and also ignore the \emph{fairness} of attributes. In this paper, we introduce concept fairness into model for first time study fairness-aware enumeration. Specifically, propose two models, called \nonesidebc~and \ntwosidebc~respectively. To efficiently enumerate all {\nonesidebc}s, present non-trivial pruning techniques, fair $\alpha$-$\beta$ core...

Graph neural networks (GNNs) have been widely adopted for semi-supervised learning on graphs. A recent study shows that the graph random network (GRAND) model can generate state-of-the-art performance this problem. However, it is difficult GRAND to handle large-scale graphs since its effectiveness relies computationally expensive data augmentation procedures. In work, we present a scalable and high-performance GNN framework GRAND+ learning. To address above issue, develop generalized forward...