An End-To-End Hyperbolic Deep Graph Convolutional Neural Network Framework

Graph Convolutional Neural Networks(GCNs)have been widely used in various fields due to their powerful capabilities in processing graph-structured data.However,GCNs encounter significant challenges when applied to scale-free graphs with power-law distributions,resulting in substantial distortions.Moreover,most of the existing GCN models are shallow structures,which restricts their ability to capture dependencies among distant nodes and more refined high-order node features in scale-free graphs with hierarchical structures.To more broadly and precisely apply GCNs to real-world graphs exhibiting scale-free or hierarchical structures and utilize multi-level aggregation of GCNs for capturing high-level information in local representations,we propose the Hyperbolic Deep Graph Convolutional Neural Network(HDGCNN),an end-to-end deep graph representation learning framework that can map scale-free graphs from Euclidean space to hyperbolic space.In HDGCNN,we define the fundamental operations of deep graph convolutional neural networks in hyperbolic space.Additionally,we introduce a hyperbolic feature transformation method based on identity mapping and a dense connection scheme based on a novel non-local message passing framework.In addition,we present a neighborhood aggregation method that combines initial structural featureswith hyperbolic attention coefficients.Through the above methods,HDGCNN effectively leverages both the structural features and node features of graph data,enabling enhanced exploration of non-local structural features and more refined node features in scale-free or hierarchical graphs.Experimental results demonstrate that HDGCNN achieves remarkable performance improvements over state-ofthe-art GCNs in node classification and link prediction tasks,even when utilizing low-dimensional embedding representations.Furthermore,when compared to shallow hyperbolic graph convolutional neural network models,HDGCNN exhibits notable advantages and performance enhancements.

Influence of rare earth Ce on hot deformation behavior of as-cast Mn18Cr18N high nitrogen austenitic stainless steel

The hot deformation behavior of Mn18Cr18N and Mn18Cr18N+Ce high nitrogen austenitic stainless steels at 1173-1473 K and 0.01-1 s^(-1) were investigated by thermal compression tests.The influence mechanism of Ce on the hot deformation behavior was analyzed by Ce-containing inclusions and segregation of Ce.The results show that after the addition of Ce,large,angular,hard,and brittle inclusions(TiN-Al_(2)O_(3),TiN,and Al_(2)O_(3)) can be modified to fine and dispersed Ce-containing inclusions(Ce-Al-O-S and TiN-Ce-Al-O-S).During the solidification,Ce-containing inclusions can be used as heterogeneous nucleation particles to refine as-cast grains.During the hot deformation,Ce-containing inclusions can pin dislocation movement and grain boundary migration,induce dynamic recrystallization(DRX)nucleation,and avoid the formation and propagation of micro cracks and gaps.In addition,during the solidification,Ce atoms enrich at the front of solid-li-quid interface,resulting in composition supercooling and refining the secondary dendrites.Similarly,during the hot deformation,Ce atoms tend to segregate at the boundaries of DRX grains,inhibiting the growth of grains.Under the synergistic effect of Ce-containing inclusions and Ce segregation,although the hot deformation resistance and hot deformation activation energy are improved,DRX is more likely to occur and the size of DRX grains is significantly refined,and the problem of hot deformation cracking can be alleviated.Finally,the microhardness of the samples was measured.The results show that compared with as-cast samples,the microhardness of hot-deformed samples increases signific-antly,and with the increase of DRX degree,the microhardness decreases continuously.In addition,Ce can affect the microhardness of Mn18Cr18N steel by affecting as-cast and hot deformation microstructures.

Effect of an external magnetic field on improved electroslag remelting cladding process

Obtaining a uniform interface temperature field plays a crucial role in the interface bonding quality of bimetal compound rolls.Therefore,this study proposes an improved electroslag remelting cladding(ESRC)process using an external magnetic field to improve the uniformity of the interface temperature of compound rolls.The improved ESRC comprises a conventional ESRC circuit and an external coil circuit.A comprehensive 3D model,including multi-physics fields,is proposed to study the effect of external magnetic fields on the multi-phys-ics fields and interface temperature uniformity.The simulated results demonstrate that the nonuniform Joule heat and flow fields cause a non-uniform interface temperature in the conventional ESRC.As for the improved ESRC,the magnetic flux density(B_(coil))along the z-axis is pro-duced by an anticlockwise current of the external coil.The rotating Lorentz force is generated from the interaction between the radial current and axial B_(coil).Therefore,the slag pool flows clockwise,which enhances circumferential effective thermal conductivity.As a result,the uniformity of the temperature field and interface temperature improve.In addition,the magnetic flux density and rotational speed of the simulated results are in good agreement with those of the experimental results,which verifies the accuracy of the improved ESRC model.Therefore,an improved ESRC is efficient for industrial production of the compound roll with a uniform interface bonding quality.