Influence Of Edges Research Articles

Dual Label-Guided Graph Refinement for Multi-View Graph Clustering

With the increase of multi-view graph data, multi-view graph clustering (MVGC) that can discover the hidden clusters without label supervision has attracted growing attention from researchers. Existing MVGC methods are often sensitive to the given graphs, especially influenced by the low quality graphs, i.e., they tend to be limited by the homophily assumption. However, the widespread real-world data hardly satisfy the homophily assumption. This gap limits the performance of existing MVGC methods on low homophilous graphs. To mitigate this limitation, our motivation is to extract high-level view-common information which is used to refine each view's graph, and reduce the influence of non-homophilous edges. To this end, we propose dual label-guided graph refinement for multi-view graph clustering (DuaLGR), to alleviate the vulnerability in facing low homophilous graphs. Specifically, DuaLGR consists of two modules named dual label-guided graph refinement module and graph encoder module. The first module is designed to extract the soft label from node features and graphs, and then learn a refinement matrix. In cooperation with the pseudo label from the second module, these graphs are refined and aggregated adaptively with different orders. Subsequently, a consensus graph can be generated in the guidance of the pseudo label. Finally, the graph encoder module encodes the consensus graph along with node features to produce the high-level pseudo label for iteratively clustering. The experimental results show the superior performance on coping with low homophilous graph data. The source code for DuaLGR is available at https://github.com/YwL-zhufeng/DuaLGR.

Scapin: Scalable Graph Structure Perturbation by Augmented Influence Maximization

Generating data perturbations to graphs has become a useful tool for analyzing the robustness of Graph Neural Networks (GNNs). However, existing model-driven methodologies can be prohibitively expensive to apply in large graphs, which hinders the understanding of GNN robustness at scale. In this paper, we present Scapin, a data-driven methodology that opens up a new perspective by connecting graph structure perturbation for GNNs with augmented influence maximization-to either facilitate desirable spreads or curtail undesirable ones by adding or deleting a small set of edges. This connection not only allows us to perform data perturbation on GNNs with computation scalability but also provides nice interpretations. To transform such connections into efficient perturbation approaches for the new GNN setting, Scapin introduces a novel edge influence model, decomposed influence maximization objectives, and a principled algorithm for edge addition by exploiting submodularity of the objectives. Empirical studies demonstrate that Scapin can give orders of magnitude improvement over state-of-art methods in terms of runtime and memory efficiency, with comparable or even better performance.

Evolution mechanism of the local network structure recorded in distribution of distances between neighbors of each vertex

The explosive development of research on real networks have led to the proposal of many mathematical network models explaining properties shared in the real networked systems. However, considering that different models can similarly reproduce the typical properties of networks, the estimation of the evolutionary process is more essential for appropriate network modeling than is the description of current network structures. In this paper, we propose a new measure “distribution of distances between neighbors (DDN) of each vertex” to investigate network development. We first examine several model networks, including growth models based on the attachment rule and on the local rule and their hybrid models, and show that the DDN profiles of these networks accurately reflect the differences in growth mechanisms. In particular, for the local rule model, the dependence of the mean DDN of vertices on the vertex degree records the evolution of the local network structure. The application of the same analysis to several real networks show that they share the characteristics of DDN profiles with the local rule and hybrid models. These results imply that these model networks appropriately describe the evolutionary processes of real networks. In addition, we examine the characteristics of DDN specific to fractal and non-fractal networks, and show that the presence or absence of fractality of some real networks can be explained by the influence of long-range edges on the local rule. An exception to this is the co-authorships network, which is expected to be generated by connecting several fractal sub-networks.

Data on vegetation across forest edges from the FERN (Forest Edge Research Network).

Many studies have focused on vegetation across forest edges to study impacts of edges created by human activities on forest structure and composition, or patterns of vegetation at inherent natural edges. Our objective was to create a database of plant-related variables across different types of edges from various studies (mainly from across Canada, but also in Brazil and Belize) to facilitate edge research. We compiled data on vegetation along more than 300 transects perpendicular to forest edges adjacent to clear-cuts, burned areas, bogs, lakes, barrens, insect disturbances, and riparian areas from 24 studies conducted over the past three decades. Data were compiled for more than 400 plant species and forest structure variables (e.g., trees, logs, canopy cover). All data were collected with a similar sampling design of quadrats along transects perpendicular to forest edges, but with varying numbers of transects and quadrats, and distances from the edge. The purpose for most of the studies was either to determine the distance of edge influence (edge width) or to explore the pattern of vegetation along the edge to interior gradient. We provide data tables for the cover of plant species and functional groups, the species and size of live and dead trees, the density of saplings, maximum height of functional groups and shrub species, and the cover of functional groups at different heights (vertical distribution of vegetation). The Forest Edge Research Network (FERN) database provides extensive data on many variables that can be used for further study including meta-analyses and can assist in answering questions important to conservation efforts (e.g., how is distance of edge influence from created edges affected by different factors?). We plan to expand this database with subsequent studies from the authors and we invite others to contribute to make this a more global database. The data are released under a CC0 license. When using these data, we ask that you cite this data paper and any relevant publications listed in our metadata file. We also encourage you to contact the first author if you are planning to use or contribute to this database.

Cross-community shortcut detection based on network representation learning and structural features

As social networks continue to expand, an increasing number of people prefer to use social networks to post their comments and express their feelings, and as a result, the information contained in social networks has grown explosively. The effective extraction of valuable information from social networks has attracted the attention of many researchers. It can mine hidden information from social networks and promote the development of social network structures. At present, many ranking node approaches, such as structural hole spanners and opinion leaders, are widely adopted to extract valuable information and knowledge. However, approaches for analyzing edge influences are seldom considered. In this study, we proposed an edge PageRank to mine shortcuts (these edges without direct mutual friends) that are located among communities and play an important role in the spread of public opinion. We first used a network-embedding algorithm to order the spanners and determine the direction of every edge. Then, we transferred the graphs of social networks into edge graphs according to the ordering. We considered the nodes and edges of the graphs of the social networks as edges and nodes of the edge graphs, respectively. Finally, we improved the PageRank algorithm on the edge graph to obtained the edge ranking and extracted the shortcuts of social networks. The experimental results for five different sizes of social networks, such as email, YouTube, DBLP-L, DBLP-M, and DBLP-S, verify whether the inferred shortcut is indeed more useful for information dissemination, and the utility of three sets of edges inferred by different methods is compared, namely, the edge inferred by ER, the edge inferred by the Jaccard index. The ER approach improves by approximately 10%, 9.9%, and 8.3% on DBLP, YouTube, and Orkut. Our method is more effective than the edge ranked by the Jaccard index.

UAV‐Lidar reveals that canopy structure mediates the influence of edge effects on forest diversity, function and microclimate

Abstract Widespread forest loss and fragmentation dramatically increases the proportion of forest areas located close to edges. Although detrimental, the precise extent and mechanisms by which edge proximity impacts remnant forests remain to be ascertained. By combining unmanned aerial vehicle laser scanning (UAV‐LS) with field data from 46 plots distributed at varying distances from the edge to the forest interior in a fragmented forest of New‐Caledonia, we investigated edge influence on forest structure, composition, function, above‐ground biomass (AGB) and microclimate. Using simple linear regressions, structural equation modelling and variance partitioning, we analysed the direct and indirect relationships between distance to edge, UAV‐LS‐derived canopy structural metrics, understorey microclimate, AGB, taxonomic and functional composition while accounting for the potential influence of fine‐scale variation in topography. We found that the distance to the closest forest edge was strongly correlated with canopy structure and that canopy structure was better correlated to forest composition, function, biomass and microclimate than distance to the closest forest edge. This suggests that the influence of edge is mediated by changes in canopy structure. Plots located near the edge exhibited a lower canopy with more gaps, higher microclimate extremes, lower biomass, lower taxonomic and functional diversity as well as denser wood and lower specific leaf area. UAV‐LS‐derived canopy structural metrics were relevant predictors of understorey microclimate, biomass and taxonomic and functional composition. Overall, the influence of topography was marginal compared to edge effects. Synthesis. Accounting for fine‐scale variation in canopy structure captured by UAV‐LS provides insights on the multiple edge impacts on key forest properties related to structure, diversity, function, biomass and microenvironmental conditions. Integrating UAV‐LS‐derived data can foster our understanding of cascading and interacting impacts of anthropogenic influence on tropical forest ecosystems and should help to improve conservation strategies and landscape management policies.

Fatigue and Progressive Damage of Thin Woven CFRP Plates Weakened by Circular Holes

BackgroundIn order to design thin-walled components it is necessary to consider the presence of holes and their effects. For high performance composite structures, this is still an issue, since usually only coupons are used in experimental observations and the influence of free edges and the hole affects the fatigue behavior mutually.ObjectiveThis work aims to find, through experimental trials, an empirical model that can be used to describe and predict the damage propagation, originating from a circular hole.MethodsA fatigue test series is performed and the damage initiation and propagation is monitored with three-dimensional digital image correlation, with which the occurring damage can be measured. Validation of the experimentally induced damage size measured with digital image correlation is performed intermediate with an in-situ measurement with active thermography and phased array ultrasonic. The novelty of this approach is that wide specimens are used, where the influence of the free edges on the hole does not occur.ResultsThe progression of the detected damage over the test reflects the applied loads, where higher loads cause larger damage. For all defined load levels a similar damage propagation is identified, allowing to establish an empirical model and fit it to the test data.ConclusionThe proposed empirical model provides a novel approach to describe and predict damage propagation originating from a circular hole in thin-walled composite plates. In addition, it is shown that the damage propagation ceases for the selected plate configuration and thus does not lead to a complete failure.

Forest edge effects on moss growth are amplified by drought.

Forest fragmentation increases the amount of edges in the landscape. Differences in wind, radiation, and vegetation structure create edge-to-interior gradients in forest microclimate, and these gradients are likely to be more pronounced during droughts and heatwaves. Although the effects of climate extremes on edge influences have potentially strong and long-lasting impacts on forest understory biodiversity, they are not well understood and are not often considered in management and landscape planning. Here we used a novel method of retrospectively quantifying growth to assess biologically relevant edge influences likely caused by microclimate using Hylocomium splendens, a moss with annual segments. We examined how spatio-temporal variation in drought across 3 years and 46 sites in central Sweden, affected the depth and magnitude of edge influences. We also investigated whether edge effects during drought were influenced by differences in forest structure. Edge effects were almost twice as strong in the drought year compared to the non-drought years, but we did not find clear evidence that they penetrated deeper into the forest in the drought year. Edge influences were also greater in areas that had fewer days with rain during the drought year. Higher levels of forest canopy cover and tree height buffered the magnitude of edge influence in times of drought. Our results demonstrate that edge effects are amplified by drought, suggesting that fragmentation effects are aggravated when droughts become more frequent and severe. Our results suggest that dense edges and buffer zones with high canopy cover can be important ways to mitigate negative drought impacts in forest edges.

Experimental and numerical evaluation of pultruded GFRP double-lap joints with different mechanical fasteners

Evaluating the structural behavior of connections is of core importance in the design of a Pultruded Glass Fiber Reinforced Polymer (PGFRP) structure. In this paper PGFRP double lap-joint specimens mechanically fastened by Through-Bolts (TB) and Hollo-Bolts (HB) were tested under uniaxial tension test in static conditions. A three-dimensional progressive Finite Element (FE) model was proposed and validated to simulate joint failure. A parametric study was conducted to investigate the influence of edge distance-to-hole diameter (e/d), number of bolts (single, double, and triple bolts), bolt diameter, width-to-hole diameter (w/d), and thickness. The model agreed well with the experimental results regarding load vs displacement, load vs strain and failure characteristics. From the result analysis, it was found that connection resistance increased with an increase in the number of bolt rows. Hence the triple HB connection with specimen ID T12H4 demonstrated an overall ultimate failure load of 57.75 kN. While the aforementioned values were approximately 26% higher when compared to TB for the same configuration.

Blob properties in I-mode and ELM-free H-mode compared to L-mode on EAST

The blob properties in I-mode and ELM-free H-mode plasmas compared to L-mode have been investigated on the EAST tokamak, including the blob detection rate size lifetime and radial velocity The blob properties in L-mode and I-mode are similar, and those in ELM-free H-mode are different to them. The blob is smaller while is larger in H-mode. The experimental blob size and velocity scalings show a good agreement with the theoretical models. The variation in blob properties during the L-I and H-L transitions, and their relations to the scrape-off layer (SOL) density, edge and SOL turbulence, and SOL collisionality are discussed. The suppression of the edge (inside the last closed flux surface) turbulence is not reflected in the blob behavior, while the blob detection rate shows a correlation with the SOL density and its low-frequency (3–50 kHz) fluctuations. In addition, the blob detection rate is found to increase with the divertor collisionality indicating a dependence of blob behavior on The differences in blob detection rates among the three operating regimes might be due to their different SOL densities and collisionalities. The investigation contributes to understanding the influences of edge and SOL plasma parameters on the blob behavior.

Influence of a Pronounced Pre-Deformation on the Attachment of Melt Droplets and the Fatigue Behavior of Laser-Cut AISI 304

Laser cutting is a suitable manufacturing method for generating complex geometries for sheet metal components. However, their cyclic load capacity is reduced compared to, for example, milled components. This is due to the influence of the laser-cut edge, whose characteristic features act as crack initiation sites, especially resolidified material in the form of burr and melt droplets. Since sheet metal components are often formed into their final geometry after cutting, another important factor influencing fatigue behavior is the effect of the forming process on the laser-cut edge. In particular, the effect of high degrees of deformation has not yet been researched in detail. Accordingly, sheets of AISI 304 were processed by laser cutting and pre-deformed. In the process, α’-martensite content was set to be comparable despite different degrees of deformation. It was found that deformation to high elongations caused a large part of the melt adhesions to fall off, but those still attaching were partially detached and thus formed an initial notch for crack initiation. This significantly lowered the fatigue strength.

Influence of wiper edge on surface integrity for end-milling of Ti-6Al-4V

Influence of wiper edge on surface integrity for end-milling of Ti-6Al-4V

Graphene-edge-supported iron dual-atom for oxygen reduction electrocatalysts.

Pyrolyzed Fe-N-C-based catalysts, particularly FeN4, are reported to show enhanced catalytic activity for some chemical reactions, particularly for the oxygen reduction reaction (ORR). Here, we present a computational study to investigate another pyrolyzed Fe-N-C-based catalyst, i.e. Fe2N6, adsorbed on graphene with special emphasis on the edges of graphene nanoribbons (both zig-zag and armchair configurations) as a candidate for Fe dual-atom catalysts (Fe-DACs). Utilizing density functional theory calculations along with microkinetic simulations, we investigate the influence of graphitic edges on the stability and ORR activity of Fe-DAC active sites. Our findings indicate that the presence of graphitic edges, particularly the zig-zag configuration, significantly lowers the formation energy of Fe-DAC active sites, making them more likely to form at the edges. Furthermore, several Fe-DAC active sites at graphitic edges exhibit exceptional ORR performance, surpassing the commonly employed FeN4 active site in SAC systems and even exceeding the benchmark Pt(111) surface. Notably, the (Fe2N6)o@z1 active site demonstrates outstanding performance in both associative and dissociative mechanisms. These results highlight the role of graphitic nanopores in enhancing the catalytic behavior of Fe-DAC active sites, providing valuable insights for designing efficient non-precious metal catalysts for ORR applications.

Influence of wiper edge on surface integrity for end-milling of Ti-6Al-4V

Influence of wiper edge on surface integrity for end-milling of Ti-6Al-4V

Experimental Design Considerations for Assessing Atmospheric Corrosion in a Marine Environment: Surrogate C1010 Steel

A rigorous assessment of marine atmospheric corrosion at a controlled NRL test site in Key West Florida was conducted. Certain factors which have been previously implicated in the literature as influencing the corrosion of engineering materials in atmospheric exposure were isolated and explored. In particular, the effect of sample size and orientation was explored. Low carbon steel (C1010) witness coupons were exposed in vertical non-sheltered, vertical sheltered, and tilted non-sheltered conditions. The effect of surface area on measured steel mass loss was also explored to identify the veracity of the so-called “edge effect”. Efforts were made to correlate meteorological atmospheric conditions (temperature, relative humidity, wind speed, wind direction, etc.) to the monthly assessment of corrosion damage. Results were assessed in terms of steel mass loss. Additive composite monthly corrosion damage tended to significantly overshoot the observed cumulative corrosion damage for samples exposed over the same period. This observation, among others presented herein, suggests that exposure of samples for less than 6 months is not an adequate predictor of long-term, natural exposure. Additionally, a smaller sample had a larger area-normalized mass loss than a larger sample. The influence of the sample edge (especially the bottom edge) was implicated in causing this difference.

Edge and Interface Resistances Create Distinct Trade-Offs When Optimizing the Microstructure of Printed van der Waals Thin-Film Transistors

Inks based on two-dimensional (2D) materials could be used to tune the properties of printed electronics while maintaining compatibility with scalable manufacturing processes. However, a very wide range of performances have been reported in printed thin-film transistors in which the 2D channel material exhibits considerable variation in microstructure. The lack of quantitative physics-based relationships between film microstructure and transistor performance limits the codesign of exfoliation, sorting, and printing processes to inefficient empirical approaches. To rationally guide the development of 2D inks and related processing, we report a gate-dependent resistor network model that establishes distinct microstructure-performance relationships created by near-edge and intersheet resistances in printed van der Waals thin-film transistors. The model is calibrated by analyzing electrical output characteristics of model transistors consisting of overlapping 2D nanosheets with varied thicknesses that are mechanically exfoliated and transferred. Kelvin probe force microscopy analysis on the model transistors leads to the discovery that the nanosheet edges, not the intersheet resistance, limit transport due to their impact on charge carrier depletion and scattering. Our model suggests that when transport in a 2D material network is limited by the near-edge resistance, the optimum nanosheet thickness is dictated by a trade-off between charged impurity screening and gate screening, and the film mobilities are more sensitive to variations in printed nanosheet density. Removal of edge states can enable the realization of higher mobilities with thinner nanosheets due to reduced junction resistances and reduced gate screening. Our analysis of the influence of nanosheet edges on the effective film mobility not only examines the prospects of extant exfoliation methods to achieve the optimum microstructure but also provides important perspectives on processes that are essential to maximizing printed film performance.

Depth of Edge Influence in a Madagascar Lowland Rainforest and Its Effects on Lemurs’ Abundance

Edge effects result from interactions between adjacent habitats, which can modify abiotic and biotic conditions and produce various negative effects on biodiversity. Given the high degree of forest fragmentation in Madagascar, understanding lemur responses to edges is a conservation priority. We aim to determine the depth of edge influence in a continuous low-land rainforest of south-eastern Madagascar and identify the response of six lemur species. We surveyed lemur abundance along nine 1 km transects from May 2015 to July 2016 totaling 112.2 km of survey effort during the day and 88.5 km at night. We characterized the habitat structure via 33 plots centered along the line transects. We used Generalized Additive Models and Generalized Linear Models to test the effect of distance from the forest edge on vegetation parameters and animal encounter rates. Edge effect on the vegetation structure can be detected up to around 100 m in terms of tree diversity and density. We found a negative edge response for Madame Fleurette’s sportive lemurs (Lepilemur fleuretae) and collared brown lemurs (Eulemur collaris), and a positive edge response for Anosy mouse lemurs (Microcebus tanosi), Southern bamboo lemurs (Hapalemur meridionalis) and Southern woolly lemurs (Avahi meridionalis). Since around half of the forested areas in Madagascar are within 100 m of forest edge, taking into account edge effect is vital when producing estimates of population sizes and informing conservation management.

Urbanization increases the range, but not the depth, of forest edge influences on Pinus sylvestris bark pH

The influence of forest edge on Pinus sylvestris bark pH was evaluated in urban and suburban pine forests in Yekaterinburg (South Taiga subzone, Central Urals, Russia). The measurements was conducted in 18 transects (rows of 6–10 sample plots arranged in pairs along a 100–260 m line perpendicular to the forest boundary); 12 transects (110 plots) were located in the urban forest and 6 transects (58 plots) in the suburban forest. All the characteristics studied (degree of urbanization, distance from forest boundary, and type and age of boundary), were found to have a significant effect on the pH of P. sylvestris bark. Bark alkalinization increased on average by 0.2 to 0.5 pH units above background along a gradient from the forest interior to the forest edge. In urban forests, the edge influence was about twice greater (0.6–0.8 pH units) than in suburban forests (0.2–0.3 pH units), while the depth ranged between 70 and 160 m. Little variation in the depth of the edge influence observed due to transect characteristics, and no differences were noted due to degree of urbanization. The range of the edge influence was cumulative over the time that had elapsed since forest fragmentation and highway construction. The results show that urban airborne dust pollution affects southern taiga pine forests to a depth of about 100 to 150 meters from forest boundary. This circumstance is important to take into account when planning the locations of the borders of forests, roads, residential and social areas.

Investigation of Material Removal Distributions and Surface Morphology Evolution in Non-Contact Ultrasonic Abrasive Machining (NUAM) of BK7 Optical Glasses.

A non-contact ultrasonic abrasive machining approach provides a potential solution to overcome the challenges of machining efficiency in the high-precision polishing of optical components. Accurately modeling the material removal distribution (removal function (RF)) and surface morphology is very important in establishing this new computer-controlled deterministic polishing technique. However, it is a challenging task due to the absence of an in-depth understanding of the evolution mechanism of the material removal distribution and the knowledge of the evolution law of the microscopic surface morphology under the complex action of ultrasonic polishing while submerged in liquid. In this study, the formation of the RF and the surface morphology were modeled by investigating the cavitation density distribution and conducting experiments. The research results showed that the material removal caused by cavitation bubble explosions was uniformly distributed across the entire working surface and had a 0.25 mm edge influence range. The flow scour removal was mainly concentrated in the high-velocity flow zone around the machining area. The roughness of the machined surface increased linearly with an increase in the amplitude and gap. Increasing the particle concentration significantly improved the material removal rate, and the generated surface exhibited better removal uniformity and lower surface roughness.

Growth Mechanism of Single-Domain Monolayer MoS2 Nanosheets on Au(111) Revealed by In Situ Microscopy: Implications for Optoelectronics Applications

The nucleation and growth of single-layer molybdenum disulfide single-domain nanosheets is investigated by in situ low-energy electron microscopy. We study the growth of micrometer-sized flakes and the correlated flattening process of the gold surface for three different elevated temperatures. Furthermore, the influence of surface step edges on the molybdenum disulfide growth process is revealed. We show that both nanosheet and underlying terrace grow simultaneously by pushing the surface step in the expansion process. Our findings point to an optimized growth procedure allowing for step-free, single-domain, single-layer islands of several micrometers in size, which is likely transferable to other transition-metal dichalcogenides (TMDs), offering a very fine degree of control over the TMD nanosheet structure and thickness.