Complex Network Structure Research Articles

Adaptive Neural Network Structure Optimization Algorithm Based on Dynamic Nodes.

Large-scale artificial neural networks have many redundant structures, making the network fall into the issue of local optimization and extended training time. Moreover, existing neural network topology optimization algorithms have the disadvantage of many calculations and complex network structure modeling. We propose a Dynamic Node-based neural network Structure optimization algorithm (DNS) to handle these issues. DNS consists of two steps: the generation step and the pruning step. In the generation step, the network generates hidden layers layer by layer until accuracy reaches the threshold. Then, the network uses a pruning algorithm based on Hebb’s rule or Pearson’s correlation for adaptation in the pruning step. In addition, we combine genetic algorithm to optimize DNS (GA-DNS). Experimental results show that compared with traditional neural network topology optimization algorithms, GA-DNS can generate neural networks with higher construction efficiency, lower structure complexity, and higher classification accuracy.

Mean-field theory of vector spin models on networks with arbitrary degree distributions

Understanding the relationship between the heterogeneous structure of complex networks and cooperative phenomena occurring on them remains a key problem in network science. Mean-field theories of spin models on networks constitute a fundamental tool to tackle this problem and a cornerstone of statistical physics, with an impressive number of applications in condensed matter, biology, and computer science. In this work we derive the mean-field equations for the equilibrium behavior of vector spin models on high-connectivity random networks with an arbitrary degree distribution and with randomly weighted links. We demonstrate that the high-connectivity limit of spin models on networks is not universal in that it depends on the full degree distribution. Such nonuniversal behavior is akin to a remarkable mechanism that leads to the breakdown of the central limit theorem when applied to the distribution of effective local fields. Traditional mean-field theories on fully-connected models, such as the Curie–Weiss, the Kuramoto, and the Sherrington–Kirkpatrick model, are only valid if the network degree distribution is highly concentrated around its mean degree. We obtain a series of results that highlight the importance of degree fluctuations to the phase diagram of mean-field spin models by focusing on the Kuramoto model of synchronization and on the Sherrington–Kirkpatrick model of spin-glasses. Numerical simulations corroborate our theoretical findings and provide compelling evidence that the present mean-field theory describes an intermediate regime of connectivity, in which the average degree c scales as a power c ∝ N b (b < 1) of the total number N ≫ 1 of spins. Our findings put forward a novel class of spin models that incorporate the effects of degree fluctuations and, at the same time, are amenable to exact analytic solutions.

Research and Application of a Fast Recovery Strategy for Complex Structure Networks Based on Sorting Multi-objective

With the development of information technology, the scale of data centers and the number of equipment are experiencing explosive growth, and the network structure of data centers has also continued to develop and mature. The more complex network structure increases the possibility of network failures and the difficulty of operation and maintenance. When an unpredictable failure occurs in the network, in order to quickly recover the network, a multi-objective complex structure network rapid recovery strategy is proposed. Under the goal of the shortest failure time, minimum reduction of safety protection, and maximum possible restoration of the original method, a restoration plan is given.

Unlocking the Changes of Phyllosphere Fungal Communities of Fishscale Bamboo (Phyllostachys heteroclada) under Rhombic-Spot Disease Stressed Conditions

As an important nonwood bioresource, fishscale bamboo (Phyllostachys heteroclada Oliver) is widely distributed in the subtropical region of China. Rhombic-spot disease, caused by Neostagonosporella sichuanensis, is one of the most serious diseases that threatens fishscale bamboo health. However, there is limited knowledge about how rhombic-spot disease influences the diversity and structures of phyllosphere fungal communities. In this study, we investigated the phyllosphere fungal communities from stems, branches, and leaves of fishscale bamboo during a rhombic-spot disease outbreak using 18S rRNA sequencing. We found that only the phyllosphere fungal community from stems was significantly affected by pathogen invasion in terms of community richness, diversity, and structure. FUNGuild analysis revealed that the major classifications of phyllosphere fungi based on trophic modes in stems, branches, and leaves changed from symbiotroph-pathotroph, no obvious dominant trophic mode, and symbiotroph to saprotroph, saprotroph–pathotroph–symbiotroph, and saprotroph–symbiotroph, respectively, after pathogen invasion. The fungal community composition of the three tissues displayed significant differences at the genus level between healthy and diseased plants. The associations among fungal species in diseased samples showed more complex co-occurrence network structures than those of healthy samples. Taken together, our results highlight the importance of plant pathological conditions for the assembly of phyllosphere fungal communities in different tissues.

Community detection via network node vector label propagation

The discovery and analysis of community structure in complex networks is a hot topic in recent years. A community detection method called NVP is proposed based on the node vector label propagation rate of network nodes. The algorithm determines the centers by looking for NGC nodes, use the concept that the inner product of the node vector in the approximate modularity is greater than 0 to divide the initial community, and then divides the final community according to the relevant content of the label propagation rate. Experimental results show that the algorithm can effectively identify the community structure of various real-world networks and computer-generated networks. In addition, this algorithm can also obtain higher NMI values than DPC, FG, LE and DCN algorithms.

Degree-Based Topological Indices of Generalized Subdivision Double-Corona Product

Graph operations play an important role in constructing complex network structures from simple graphs. Computation of topological indices of these complex structures via graph products is an important task. In this paper, we generalized the concept of subdivision double-corona product of graphs and investigated the exact expressions of the first and second Zagreb indices, first reformulated Zagreb index, and forgotten topological index (F-index) of this graph operation.

Influence of new organic alkali activators on microstructure and strength of fly ash geopolymer

The fly ash geopolymer with improved mechanical properties was prepared by a new mixture alkali activator. In this paper, sodium tert-butanol, an organic strong alkali was used as an activator for preparing fly ash geopolymer to improve their mechanical properties. The effect of activator content and type on the macroscopic level of fly ash geopolymer was investigated experimentally by three types of activators: sodium tert-butanol, sodium silicate, and sodium tert-butoxide/sodium silicate mix activator. The microstructure of the fly ash geopolymer was characterized by ATR-FTIR, SEM-EDS, XRD, and Brunauer-Emmett-Teller (BET) physical adsorption method. The results showed that the new mixture alkali activator prepared by 5% sodium tert-butoxide and 10% sodium silicate improved the denseness and integrity of the microstructure of the fly ash geopolymer. Consequently, the mechanical properties of fly ash geopolymer are improved. The microscopic results demonstrated that the C–OH in tert-butanol after the hydrolysis of sodium tert-butoxide and Si–OH in the geopolymer can form C–O–Si bonds, forming a more complex three-dimensional network structure. This paper reveals the enhancement mechanism of organic alkali as activators for preparing fly ash geopolymer, and provides support for the subsequent development of organic strong alkali activators.

How does the position in business group hierarchies affect workers’ wages?

We merge firm-level data on ownership linkages with administrative data on German workers to analyze how the position in a business group hierarchy affects workers’ wages. To acknowledge that ownership linkages are not one-directional, we propose an index of hierarchical distance to the ultimate owner that accounts for the complex network structure of business groups. After controlling for unobserved heterogeneity, we find a positive effect of larger hierarchical distance to the ultimate owner of a business group on workers’ wages. To explain this finding, we develop a monitoring-based theory of business groups. Our model predicts higher wages to prevent shirking by workers if a larger hierarchical distance to the ultimate owner is associated with lower monitoring efficiency.

Dynamic ensemble selection of convolutional neural networks and its application in flower classification

In recent years, convolutional neural networks (CNNs) have achieved great success in image classification. However, CNN models usually have complex network structures that tend to cause some related problems, such as redundancy of network parameters, low training efficiency, overfitting, and weak generalization ability. To solve these problems and improve the accuracy of flower classification, the advantages of CNNs were combined with those of ensemble learning and a method was developed for the dynamic ensemble selection of CNNs. First, MobileNet models pre-trained on a public dataset were transferred to flower datasets to train thirteen different MobileNet classifiers, and a resampling strategy was used to enhance the diversity of individual models. Second, the thirteen classifiers were sorted by a classifier sorting algorithm, before ensemble selection, to avoid an exhaustive search. Finally, with the credibility of recognition results, a classifier subset was dynamically selected and integrated to identify the flower species from their images. To verify the effectiveness, the proposed method was used to classify the images of five flower species. The accuracy of the proposed method was 95.50%, an improvement of 1.62%, 3.94%, 22.04%, 13.77%, and 0.44%, over those of MobileNet, Inception-v1, ResNet-50, Inception-ResNet-v2, and the linear ensemble method, respectively. In addition, the performance of the proposed method was compared with five other methods for flower classification. The experimental results demonstrated the accuracy and robustness of the proposed method. Keywords: flowers, classification, convolutional neural network, dynamic ensemble selection DOI: 10.25165/j.ijabe.20221501.6313 Citation: Wang Z B, Wang K Y, Wang X F, Pan S H, Qiao X J. Dynamic ensemble selection of convolutional neural networks and its application in flower classification. Int J Agric & Biol Eng, 2022; 15(1): 216–223.

Effect of TiO2 substituting SiO2 on the rheological and crystallization behavior of mold slags for casting Ti-containing steel

In the current study, the influence of TiO 2 substituting SiO 2 on the rheological and crystallization behavior of mold slags for casting Ti-containing steel was investigated through the measurement of the viscosity-temperature relationship, crystallization process and initial crystallization temperature. The structure of molten slags from the Raman spectrum and other parameters (NBO/T and Optical basicity) was compared to explain the difference of slag properties. With the substituting of SiO 2 by TiO 2 in slags from 5% to 20%, the viscosity of these slags at 1573 K decreased from 0.36 Pa s to 0.10 Pa s, and the viscous flow activation energy below the break temperature had the tendency to decrease. With the increase of the TiO 2 /SiO 2 ratio from 0.25 to 4.00, the initial crystallization temperature of slags increased and the crystallization of slags were enhanced, which was related to the difference of the crystallization mode and the precipitation of dispersed crystals as nuclei. With the substituting of SiO 2 by TiO 2 , the parameter NBO/T and optical basicity of molten slags increased, the complex network structure by the Raman spectra depolymerized gradually and the degree of polymerization decreased, which was related to the destruction of silicate/aluminate structure by the addition of TiO 2 in slags. Hence, the substitution of SiO 2 by TiO 2 in slags simplified the network structure of mold slags, then lowered the viscosity and promoted the crystallization of slags.

TPE: Lightweight Transformer Photo Enhancement Based on Curve Adjustment

In recent years, learning-based methods have made great progress in the field of photo enhancement. However, the enhancement methods rely on complex network structures and consume excessive amounts of computing resources, which greatly increases the difficulty of their deployment on lightweight devices. Additionally, the methods have poor real-time performance when processing very large resolution images. In contrast to previous works on designing structurally diverse CNN networks, photo enhancement can be achieved through a lightweight self-attentive mechanism for global-local tuning. In this paper, we design a lightweight photo enhancement tool based on Transformer; we dub the tool, TPE. TPE captures long-range dependencies among image patches and can efficiently extract the structural relationships within an image. A multistage curve adjustment strategy overcomes the problem of the limited adjustment capabilities of the global adjustment function, allowing the method to combine both global modifications and local fine-tuning. Experiments on various benchmarks demonstrate the qualitative and quantitative advantages of TPE over state-of-the-art methods in photo retouching and low-light image enhancement tasks.

Forecasting Fine-Grained Urban Flows Via Spatio-Temporal Contrastive Self-Supervision

As a critical task of the urban traffic services, fine-grained urban flow inference (FUFI) benefits in many fields including intelligent transportation management, urban planning, public safety. FUFI is a technique that focuses on inferring fine-grained urban flows depending solely on observed coarse-grained data. However, existing methods always require massive learnable parameters and the complex network structures. To reduce these defects, we formulate a contrastive self-supervision method to predict fine-grained urban flows taking into account all correlated spatial and temporal contrastive patterns. Through several well-designed self-supervised tasks, uncomplicated networks have a strong ability to capture high-level representations from flow data. Then, a fine-tuning network combining with three pre-training encoder networks is proposed. We conduct experiments to evaluate our model and compare with other state-of-the-art methods by using two real-world datasets. All the empirical results not only show the superiority of our model against other comparative models, but also demonstrate its effectiveness in the resource-limited environment.

A Modular Framework for Centrality and Clustering in Complex Networks

The structure of many complex networks includes edge directionality and weights on top of their topology. Network analysis that can seamlessly consider combination of these properties are desirable. In this paper, we study two important such network analysis techniques, namely, centrality and clustering. An information-flow based model is adopted for clustering, which itself builds upon an information theoretic measure for computing centrality. Our principal contributions include (1) a generalized model of Markov entropic centrality with the flexibility to tune the importance of node degrees, edge weights and directions, with a closed-form asymptotic analysis, which (2) leads to a novel two-stage graph clustering algorithm. The centrality analysis helps reason about the suitability of our approach to cluster a given graph, and determine ‘query’ nodes, around which to explore local community structures, leading to an agglomerative clustering mechanism. Our clustering algorithm naturally inherits the flexibility to accommodate edge directionality, as well as different interpretations and interplay between edge weights and node degrees. Extensive benchmarking experiments are provided, using both real-world networks with ground truth and synthetic networks.

Microscopic Structural Analysis of Complex Networks: An Empirical Study Using Motifs

Complex Networks can depict a clear image of real-world systems. A real-world scenario can be represented a graph with interconnected layers - called a multilayer network. Finding motifs can give an idea of the topology of complex systems and help understand the graphs&#x2019; dynamics. Looking at motifs as atoms of the network is helpful to analyze the relationship between nodes and between layers. This work suggests a sub-graph enumeration approach to find and count the motifs in a multilayer network. The proposed work has many applications in graph mining, particularly to the structure and dynamics of complex networks.

Global Transfer of Salinization on Irrigated Land: Complex Network and Endogenous Structure

Global Transfer of Salinization on Irrigated Land: Complex Network and Endogenous Structure

Enhanced Spectral–Spatial Residual Attention Network for Hyperspectral Image Classification

Deep learning has achieved good performance in hyperspectral image classification (HSIC). Many methods based on deep learning use deep and complex network structures to extract rich spectral and spatial features of hyperspectral images (HSIs) with high accuracy. During the process, how to accurately extract the features and information from pixel blocks in HSIs is important. All of the spectral features are treated equally in classification, and the input of the network often contains much useless pixel information, leading to a low classification result. To solve this problem, an enhanced spectral-spatial residual attention network (ESSRAN) is proposed for HSIC in this paper. In the proposed network, the spectral-spatial attention network (SSAN), residual network (ResNet) and long-short term memory (LSTM) are combined to extract more discriminative spectral and spatial features. More specifically, SSAN is first applied to extract image features by using the spectral attention module to emphasize useful bands and suppress useless bands. The spatial attention module is used to emphasize pixels that have same category with the central pixel. Then, these obtained features are fed into an improved ResNet, which adopts LSTM to learn representative high-level semantic features of the spectral sequences, since the use of ResNet can prevent gradient disappearance and explosion. The proposed ESSRAN model is implemented on three commonly used HSI datasets and compared to some state-of-the-art methods. The results confirm that ESSRAN effectively improves accuracy.

Invulnerability Analysis and Optimization Strategy of Sector Network Using Cascading Failure Model

Resolving the challenge of flight delays caused by air traffic congestion renders it necessary to explore the mode of congestion propagation. By applying complex network theory, this article establishes a complex network structure where airspace sectors act as nodes, and the edges represent traffic flow relationships between sectors. In addition, a cascading failure model is proposed to analyze the airspace sector network’s invulnerability. The critical threshold and the sector abnormality rate are determined according to the cascading failure measurement indexes. Based on the remaining capacity of sector nodes, two optimization strategies are proposed: adjacent load redistribution strategy and local load redistribution strategy. The simulation results demonstrate that the local load redistribution strategy greatly improves the airspace sector network’s invulnerability to cascading failure.

SETR-YOLOv5n: A Lightweight Low-Light Lane Curvature Detection Method Based on Fractional-Order Fusion Model

End-to-end automatic driving requires the identification of lane curvature. We proposed a lightweight detection method for the low-light lane curvature based on the Fractional-Order Fusion Model (FFM) to assure real-time performance and increase the reliability of automatic driving in low-light conditions. To begin, the FFM method is introduced to enhance images with low average brightness, fuzzy detail, and a high signal-to-noise ratio. Under low-light conditions, these images cannot clearly express information (such as rainy, snowy, foggy, and other harsh external environments). Then, aiming at the problems of complex network structure, high hardware configuration required for training, and low transmission Frames Per Second (FPS) of real-time detection in the previously proposed YOLOv5, the SETR-C3Block module is proposed. The YOLOv5n is improved by optimizing the configuration of the target detector head and the network’s structure, which solves the problems of low efficiency and redundancy parameters in feature extraction in the network. According to the experimental results on the lane curvature dataset, the mAP@.5:.95 of SETR-YOLOv5n is 87.22%, the transmission FPS of real-time detection is 70.4, and the number of model parameters is only 1.8M. It shows that the SETR-YOLOv5n can meet the lightweight and accuracy requirements of target detection by the mobile terminal or embedded device.

MAPPINGS, a tool for network analysis of large phospho-signalling datasets: application to host erythrocyte response to Plasmodium infection

Large datasets of phosphorylation interactions are constantly being generated, but deciphering the complex network structure hidden in these datasets remains challenging. Many phosphorylation interactions occurring in human cells have been identified and constitute the basis for the known phosphorylation interaction network. We overlayed onto this network phosphorylation datasets obtained from an antibody microarray approach aimed at determining changes in phospho-signalling of host erythrocytes, during infection with the malaria parasite Plasmodium falciparum. We designed a pathway analysis tool denoted MAPPINGS that uses random walks to identify chains of phosphorylation events occurring much more or much less frequently than expected. MAPPINGS highlights pathways of phosphorylation that work synergistically, providing a rapid interpretation of the most critical pathways in each dataset. MAPPINGS confirmed several signalling interactions previously shown to be modulated by infection, and revealed additional interactions which could form the basis of numerous future studies. The MAPPINGS analysis strategy described here is widely applicable to comparative phosphorylation datasets in any context, such as response of cells to infection, treatment, or comparison between differentiation stages of any cellular population.

Bidirection-Fusion-YOLOv3: An Improved Method for Insulator Defect Detection Using UAV Image

With advances in the continuous improvement and development of the power system, insulators have gradually become one of the most important components. At present, unmanned aerial vehicles(UAV) have been widely used to inspect insulators, and insulators in pictures are accurately and efficiently identified by convolutional neural networks and this method has been extensively applied. These existing methods have been widely used to identify insulators in pictures, with high accuracy and efficiency. However, they are based on Faster R-CNN and you only look once(YOLO) either require more identification time due to the complex network structure, or do not have sufficient accuracy for insulator defects. More identification time is required due to the complexity of the network structure, or there is not enough accuracy for insulator defects. Based on the YOLOv3 network, this paper proposes a new type of convolutional neural network(CNN) for target detection, which can improve enhance the efficiency while ensuring the detection speed. In addition, this paper applies the latest EIoU and loss functions to YOLOv3, which significantly improves the coincidence of the prediction frame and the annotation frame and accelerates the convergence speed. The experimental results show that the detection model proposed in this paper has an average precision(AP) of 0.94 for insulators and 0.89 for insulator defects, and its detection speed can reach 93.5ms/image. Finally, after experimental verification, the detection model proposed in this paper meets the requirements of power inspection and has good engineering application prospects.