Search Results (329)

Graph anomaly detection is crucial in many high-impact applications across diverse fields. In anomaly detection tasks, collecting plenty of annotated data tends to be costly and laborious. As a result, few-shot learning has been explored to address the issue by requiring only a few labeled samples to achieve good performance. However, conventional few-shot models may not fully exploit the information within auxiliary sets, leading to suboptimal performance. To tackle these limitations, we propose a dual-level knowledge distillation-based approach for graph anomaly detection, DualKD, which leverages two distinct distillation losses to improve generalization capabilities. In our approach, we initially train a teacher model to generate prediction distributions as soft labels, capturing the entropy of uncertainty in the data. These soft labels are then employed to construct the corresponding loss for training a student model, which can capture more detailed node features. In addition, we introduce two representation distillation losses—short and long representation distillation—to effectively transfer knowledge from the auxiliary set to the target set. Comprehensive experiments conducted on four datasets verify that DualKD remarkably outperforms the advanced baselines, highlighting its effectiveness in enhancing identification performance. Full article

Knowledge graphs, as an important research direction in artificial intelligence, have been widely applied in many fields and tasks. The relations in knowledge graphs have explicit semantics and play a crucial role in knowledge completion and reasoning. Correctly measuring the inferential value of relations and identifying important relations in a knowledge graph can effectively improve the effectiveness of knowledge graphs in reasoning tasks. However, the existing methods primarily consider the connectivity and structural characteristics of relations, but neglect the semantics and the mutual influence of relations in reasoning tasks. This leads to truly valuable relations being difficult to fully utilize in long-chain reasoning. To address this problem, this work, inspired by information entropy and uncertainty-measurement methods in knowledge bases, proposes a method called Relation Importance Measurement based on Information Entropy (RIMIE) to measure the inferential value of relations in knowledge graphs. RIMIE considers the semantics of relations and the role of relations in reasoning. Specifically, based on the values of relations in logical chains, RIMIE partitions the logical sample set into multiple equivalence classes, and generates a knowledge structure for each relation. Correspondingly, to effectively measure the inferential values of relations in knowledge graphs, the concept of relation entropy is proposed, and it is calculated according to the knowledge structures. Finally, to objectively assess the effectiveness of RIMIE, a group of experiments are conducted, which compare the influences of the relations selected according to RIMIE and other patterns on the triple classifications by knowledge graph representation learning. The experimental results confirm what is claimed above. Full article

The number and types of phases formed in high entropy alloys (HEAs) have significant impacts on the mechanical properties. While various machine learning approaches were developed for predicting whether an HEA is single or multiphase, changes in chemistry and/or composition can lead to other changes across length scales, which affect material performance. To address this challenge, we introduce a graph-based approach, which captures the similarity of alloys across these length scales, and which defines design pathways for the chemical modifications of alloys. Our network defines different regimes of alloys and therefore allows one to design within the same material regime. This approach, which also provides a new genre of HEA phase diagrams, enhances the design of alloys through control of the phase(s) present while maintaining other relevant alloy properties. Full article

An algebraic graph is defined in terms of graph theory as a graph with related algebraic structures or characteristics. If the vertex set of a graph G is a group, a ring, or a field, then G is called an algebraic structure graph. This work uses an algebraic structure graph based on the modular ring Zn, known as a hyper-chordal ring network. The lower and upper bounds of the local fractional metric dimension are computed for certain families of hyper-chordal ring networks. Utilizing the cardinalities of local fractional resolving sets, local fractional resolving (LFR)M-polynomials are computed for hyper-chordal ring networks. Further, new topological indices based on (LFR)M-polynomials are established for the proposed networks. The local fraction entropies are developed by modifying the first three kinds of Zagreb entropies, which are calculated for the chosen hyper-chordal ring networks. Furthermore, numerical and graphical comparisons are discussed to observe the order between newly computed topological indices. Full article

To develop a rapid, reliable, and cost-effective method for predicting the structure of single-phase high-entropy alloys, a Graph Neural Network (ALIGNN-FF)-based approach was introduced. This method was successfully tested on 132 different high-entropy alloys, and the results were analyzed and compared with density functional theory and valence electron concentration calculations. Additionally, the effects of various factors on prediction accuracy, including lattice parameters and the number of supercells with unique atomic configurations, were investigated. The ALIGNN-FF-based approach was subsequently used to predict the structure of a novel cobalt-free 3d high-entropy alloy, and the result was experimentally verified. Full article

Traffic congestion remains a significant challenge in urban management, with traditional fixed-cycle traffic signal systems struggling to adapt to dynamic traffic conditions. This paper proposes an adaptive traffic signal control method based on a Graph Neural Network (GNN) and a dynamic entropy-constrained Soft Actor–Critic (DESAC) algorithm. The approach first extracts both global and local features of the traffic network using GNN and then utilizes the DESAC algorithm to optimize traffic signal control at both single and multi-intersection levels. Finally, a simulation environment is established on the CityFlow platform to evaluate the proposed method’s performance through experiments involving single and twelve intersection scenarios. Simulation results on the CityFlow platform demonstrate that G-DESAC significantly improves traffic flow, reduces delays and queue lengths, and enhances intersection capacity compared to other algorithms. In single intersection scenarios, G-DESAC achieves a higher reward, reduced total delay time, minimized queue lengths, and improved throughput. In multi-intersection scenarios, G-DESAC maintains high rewards with stable and efficient optimization, outperforming DQN, SAC, Max-Pressure, and DDPG. This research highlights the potential of deep reinforcement learning (DRL) in urban traffic management and positions G-DESAC as a robust solution for practical traffic signal control applications, offering substantial improvements in traffic efficiency and congestion mitigation. Full article

The processes of information exchange and the movement of material flows form a communication space that reflects the relationship of complex intersystem interactions in various spheres of our life within the framework of the concepts of information-theoretical theory. One of these concepts, reflecting the mutual influence between processes at a qualitative level, is the transfer of entropy. The direction and intensity of these flows reflect the main social and economic processes. As it is known, air transport is one of the most reliable and high-speed modes of transport, influencing the processes of socio-cultural interaction between different regions. This indirectly affects the development of industrial relations, the development of technology and intercultural exchange. New technologies in aviation improve the flight performance of airliners and reduce the costs of transporting passengers. The size and range of modern airliners are increasing, and ticket prices are being optimized. The processes of the liberalization of developing air transportation markets, the emergence of low-cost air carriers, open skies agreements, and the reduction in restrictions on the nomenclature of carriers and routes have led to the growth and diversity of air transport links. This article considers air transport as a complex system that takes into account the interconnectedness of the elements of the transportation system and the influence of some subsystems on others, which are not always obvious. The object of the study was the communication space formed on the basis of air transportation between regions of the world. To assess the dynamic properties of the world communication space, ICAO data for the period of 1970–2021 were used. The subject of the analysis was a time series reflecting the flows of passengers and cargo over the considered time horizon. The entropy transfer algorithm was used as an analysis tool. In the course of the research, the features of dynamic changes in the properties of the communication space were revealed. The analysis showed that the flows of entropy transfer between regions of the world change depending on political, economic, social, and technological factors. Examples of the application of the proposed approach are considered: an analysis of the cognitive model of the air transport flow structure, an analysis of the regional communication space, and an analysis of changes in the global communication field. The results of the analysis can be useful for assessing the development of the communication field of various regions, which will allow us to solve the problems of forming forecasts and effective scenarios for the development of transport flows at different hierarchical levels of economic management. Full article

Due to the increasing integration of new energy sources, the power system now exhibits low inertia, in which the broadband oscillation problem is increasingly significant in the face of the strong coupling of complex and variable power systems, and the current lack of uniform and effective mathematical models and analysis methods. To solve this major problem, a broadband oscillation localization method based on the combination of compressed perception and graph attention network (GAT) is proposed. The method firstly uses the principle of compression perception to compress and transmit the oscillation time series data of the sub-station, reconstructs the compressed signal at the master station and aggregates the grid topology and node characteristic information to effectively reduce the redundancy of the oscillation data; reconstruction error is only 0.031, takes into account the balance of the samples and the effectiveness of the computation, and adopts the multi-attention mechanism and the cross-entropy loss function to improve the performance of the model training. Finally, the offline training and online evaluation model based on the GAT algorithm is constructed, and the accuracy of the model is up to 98.5%; and the results show that the method has a high positioning accuracy and a certain anti-noise ability at the same time. Full article

Rolling bearings, as critical components of rotating machinery, significantly influence equipment reliability and operational efficiency. Accurate fault diagnosis is therefore crucial for maintaining industrial production safety and continuity. This paper presents a new fault diagnosis method based on FCEEMD multi-complexity low-dimensional features and directed acyclic graph LSTSVM. The Fast Complementary Ensemble Empirical Mode Decomposition (FCEEMD) method is applied to decompose vibration signals, effectively reducing background noise. Nonlinear complexity features are then extracted, including sample entropy (SE), permutation entropy (PE), dispersion entropy (DE), Gini coefficient, the square envelope Gini coefficient (SEGI), and the square envelope spectral Gini coefficient (SESGI), enhancing the capture of the signal complexity. In addition, 16 time-domain and 13 frequency-domain features are used to characterize the signal, forming a high-dimensional feature matrix. Robust unsupervised feature selection with local preservation (RULSP) is employed to identify low-dimensional sensitive features. Finally, a multi-classifier based on DAG LSTSVM is constructed using the directed acyclic graph (DAG) strategy, improving fault diagnosis precision. Experiments on both laboratory bearing faults and industrial check valve faults demonstrate nearly 100% diagnostic accuracy, highlighting the method’s effectiveness and potential. Full article

The primary objective of our study is to analyze how the nature of explanatory variables influences the values and behavior of impurity measures, including the Shannon, Rényi, Tsallis, Sharma–Mittal, Sharma–Taneja, and Kapur entropies. Our analysis aims to use these measures in the interactive learning of decision trees, particularly in the tie-breaking situations where an expert needs to make a decision. We simulate the values of explanatory variables from various probability distributions in order to consider a wide range of variability and properties. These probability distributions include the normal, Cauchy, uniform, exponential, and two beta distributions. This research assumes that the values of the binary responses are generated from the logistic regression model. All of the six mentioned probability distributions of the explanatory variables are presented in the same graphical format. The first two graphs depict histograms of the explanatory variables values and their corresponding probabilities generated by a particular model. The remaining graphs present distinct impurity measures with different parameters. In order to examine and discuss the behavior of the obtained results, we conduct a sensitivity analysis of the algorithms with regard to the entropy parameter values. We also demonstrate how certain explanatory variables affect the process of interactive tree learning. Full article

Urban vitality is a concept that reflects dynamic changes in economic, social, and cultural aspects, emphasizing the importance of diversified land use and dense population. With technological advancements, research methods on urban vitality are increasingly diverse, particularly with the application of big data and geographic information systems providing new perspectives and tools for such studies. Currently, research on the vitality of inland Central Plains cities in China is relatively limited and largely confined to specific administrative areas, leading to an inadequate understanding of basic economic activities and population distribution within cities. Therefore, this study aims to explore the spatial distribution characteristics of urban vitality and its influencing factors in Zhengzhou’s main urban area, providing a scientific basis for urban planning and sustainable development. This study utilizes methods that include Densi graph curve analysis, the entropy method, and the multiscale geographically weighted regression (MGWR) model, integrating statistical data, geographic information, and remote sensing imagery of Zhengzhou in 2023. The MGWR model analysis reveals: (1) Urban vitality in Zhengzhou’s main urban area exhibits a concentric pattern, with high vitality at the center gradually decreasing toward the periphery, showing significant spatial differences in economic, population, and cultural vitality. (2) Various influencing factors positively correlate with urban vitality in the main urban area, but due to shortcomings in urban development strategies and planning, some factors negatively impact vitality in the central area while positively affecting vitality in peripheral areas. Based on these findings, this study provides relevant evidence and theoretical support for urban planning and sustainable development in Zhengzhou, aiding in the formulation of more effective urban development strategies. Full article

Research on named entity recognition (NER) and command-line generation for network security evaluation tools is relatively scarce, and no mature models for recognition or generation have been developed thus far. Therefore, in this study, the aim is to build a specialized corpus for network security evaluation tools by combining knowledge graphs and information entropy for automatic entity annotation. Additionally, a novel NER approach based on the KG-BERT-BiLSTM-CRF model is proposed. Compared to the traditional BERT-BiLSTM model, the KG-BERT-BiLSTM-CRF model demonstrates superior performance when applied to the specialized corpus of network security evaluation tools. The graph attention network (GAT) component effectively extracts relevant sequential content from datasets in the network security evaluation domain. The fusion layer then concatenates the feature sequences from the GAT and BiLSTM layers, enhancing the training process. Upon successful NER execution, in this study, the identified entities are mapped to pre-established command-line data for network security evaluation tools, achieving automatic conversion from textual content to evaluation commands. This process not only improves the efficiency and accuracy of command generation but also provides practical value for the development and optimization of network security evaluation tools. This approach enables the more precise automatic generation of evaluation commands tailored to specific security threats, thereby enhancing the timeliness and effectiveness of cybersecurity defenses. Full article

Deep hashing technology, known for its low-cost storage and rapid retrieval, has become a focal point in cross-modal retrieval research as multimodal data continue to grow. However, existing supervised methods often overlook noisy labels and multiscale features in different modal datasets, leading to higher information entropy in the generated hash codes and features, which reduces retrieval performance. The variation in text annotation information across datasets further increases the information entropy during text feature extraction, resulting in suboptimal outcomes. Consequently, reducing the information entropy in text feature extraction, supplementing text feature information, and enhancing the retrieval efficiency of large-scale media data are critical challenges in cross-modal retrieval research. To tackle these, this paper introduces the Text-Enhanced Graph Attention Hashing for Cross-Modal Retrieval (TEGAH) framework. TEGAH incorporates a deep text feature extraction network and a multiscale label region fusion network to minimize information entropy and optimize feature extraction. Additionally, a Graph-Attention-based modal feature fusion network is designed to efficiently integrate multimodal information, enhance the affinity of the network for different modes, and retain more semantic information. Extensive experiments on three multilabel datasets demonstrate that the TEGAH framework significantly outperforms state-of-the-art cross-modal hashing methods. Full article

When single-phase ground faults occur in distribution systems, the fault characteristics of zero-sequence current signals are not prominent. They are quickly submerged in noise, leading to difficulties in fault section location. This paper proposes a method for fault section location in distribution networks based on improved empirical wavelet transform (IEWT) and GINs to address this issue. Firstly, based on kurtosis, EWT is optimized using the N-point search method to decompose the zero-sequence current signal into modal components. Noise is filtered out through weighted permutation entropy (WPE), and signal reconstruction is performed to obtain the denoised zero-sequence current signal. Subsequently, GINs are employed for graph classification tasks. According to the topology of the distribution network, the corresponding graph is constructed as the input to the GIN. The denoised zero-sequence current signal is the node input for the GIN. The GIN autonomously explores the features of each graph structure to achieve fault section location. The experimental results demonstrate that this method has strong noise resistance, with a fault section location accuracy of up to 99.95%, effectively completing fault section location in distribution networks. Full article