Heterogeneous Hyperedge Convolutional Network

Graph convolutional networks (GCNs) have been developed as a general and powerful tool to handle various tasks related to graph data. However, current methods mainly consider homogeneous networks and ignore the rich semantics and multiple types of objects that are common in heterogeneous information networks (HINs). In this paper, we present a Heterogeneous Hyperedge Convolutional Network (HHCN), a novel graph convolutional network architecture that operates on HINs. Specifically, we extract the rich semantics by different metastructures and adopt hyperedge to model the interactions among metastructure-based neighbors. Due to the powerful information extraction capabilities of metastructure and hyperedge, HHCN has the flexibility to model the complex relationships in HINs by setting different combinations of metastructures and hyperedges. Moreover, a metastructure attention layer is also designed to allow each node to select the metastructures based on their importance and provide potential interpretability for graph analysis. As a result, HHCN can encode node features, metastructure-based semantics and hyperedge information simultaneously by aggregating features from metastructure-based neighbors in a hierarchical manner. We evaluate HHCN by applying it to the semi-supervised node classification task. Experimental results show that HHCN outperforms state-of-the-art graph embedding models and recently proposed graph convolutional network models.     

Identification of Anomaly Scenes in Videos Using Graph Neural Networks

Generally, conventional methods for anomaly detection rely on clustering, proximity, or classification. With the massive growth in surveillance videos, outliers or anomalies find ingenious ways to obscure themselves in the network and make conventional techniques inefficient. This research explores the structure of Graph neural networks (GNNs) that generalize deep learning frameworks to graph-structured data. Every node in the graph structure is labeled and anomalies, represented by unlabeled nodes, are predicted by performing random walks on the node-based graph structures. Due to their strong learning abilities, GNNs gained popularity in various domains such as natural language processing, social network analytics and healthcare. Anomaly detection is a challenging task in computer vision but the proposed algorithm using GNNs efficiently performs the identification of anomalies. The Graph-based deep learning networks are designed to predict unknown objects and outliers. In our case, they detect unusual objects in the form of malicious nodes. The edges between nodes represent a relationship of nodes among each other. In case of anomaly, such as the bike rider in Pedestrians data, the rider node has a negative value for the edge and it is identified as an anomaly. The encoding and decoding layers are crucial for determining how statistical measurements affect anomaly identification and for correcting the graph path to the best possible outcome. Results show that the proposed framework is a step ahead of the traditional approaches in detecting unusual activities, which shows a huge potential in automatically monitoring surveillance videos. Performing autonomous monitoring of CCTV, crime control and damage or destruction by a group of people or crowd can be identified and alarms may be triggered in unusual activities in streets or public places. The suggested GNN model improves accuracy by 4% for the Pedestrian 2 dataset and 12% for the Pedestrian 1 dataset compared to a few state-of-the-art techniques.     

Advanced Community Identification Model for Social Networks

Community detection in social networks is a hard problem because of the size, and the need of a deep understanding of network structure and functions. While several methods with significant effort in this direction have been devised, an outstanding open problem is the unknown number of communities, it is generally believed that the role of influential nodes that are surrounded by neighbors is very important. In addition, the similarity among nodes inside the same cluster is greater than among nodes from other clusters. Lately, the global and local methods of community detection have been getting more attention. Therefore, in this study, we propose an advanced community-detection model for social networks in order to identify network communities based on global and local information. Our proposed model initially detects the most influential nodes by using an Eigen score then performs local expansion powered by label propagation. This process is conducted with the same color till nodes reach maximum similarity. Finally, the communities are formed, and a clear community graph is displayed to the user. Our proposed model is completely parameter-free, and therefore, no prior information is required, such as the number of communities, etc. We perform simulations and experiments using well-known synthetic and real network benchmarks, and compare them with well-known state-of-the-art models. The results prove that our model is efficient in all aspects, because it quickly identifies communities in the network. Moreover, it can easily be used for friendship recommendations or in business recommendation systems.     

MINE: A Method of Multi-Interaction Heterogeneous Information Network Embedding

Interactivity is the most significant feature of network data, especially in social networks. Existing network embedding methods have achieved remarkable results in learning network structure and node attributes, but do not pay attention to the multiinteraction between nodes, which limits the extraction and mining of potential deep interactions between nodes. To tackle the problem, we propose a method called MultiInteraction heterogeneous information Network Embedding (MINE). Firstly, we introduced the multi-interactions heterogeneous information network and extracted complex heterogeneous relation sequences by the multi-interaction extraction algorithm. Secondly, we use a well-designed multi-relationship network fusion model based on the attention mechanism to fuse multiple interactional relationships. Finally, applying a multitasking model makes the learned vector contain richer semantic relationships. A large number of practical experiments prove that our proposed method outperforms existing methods on multiple data sets.     

基于VMD-ICMSE和半监督判别SOINN L-Isomap的滚动轴承故障诊断

针对从滚动轴承非线性、非平稳振动信号中提取故障特征困难的问题,提出一种基于半监督判别自组织增量学习神经网络界标点的等度规映射(SSDSL-Isomap)的滚动轴承故障诊断方法。利用基于变分模态分解的改进复合多尺度样本熵(VMD-ICMSE)从复杂域提取振动信号的故障特征,构建高维故障特征集;采用SSDSL-Isomap方法对高维故障特征集进行维数约简,提取出利于识别的低维、敏感故障特征子集;应用粒子群优化极限学习机(PSO-ELM)分类器对低维故障特征进行故障识别,判别故障类型。VMD-ICMSE方法集成了VMD自适应分解非线性信号与ICMSE衡量时间序列复杂性程度的优势,提高故障特征提取能力;SSDSL-Isomap方法综合了全局流形结构、半监督型双约束图构建以及SOINN界标点选取的优点,增强故障分类能力。调心球轴承故障诊断实验分析结果表明,该方法对实验数据的故障识别率达到100%。     

An Erebus Attack Detection Method Oriented to Blockchain Network Layer

Recently, the Erebus attack has proved to be a security threat to the blockchain network layer, and the existing research has faced challenges in detecting the Erebus attack on the blockchain network layer. The cloud-based active defense and one-sidedness detection strategies are the hindrances in detecting Erebus attacks. This study designs a detection approach by establishing a ReliefF_WMRmR-based two-stage feature selection algorithm and a deep learning-based multimodal classification detection model for Erebus attacks and responding to security threats to the blockchain network layer. The goal is to improve the performance of Erebus attack detection methods, by combining the traffic behavior with the routing status based on multimodal deep feature learning. The traffic behavior and routing status were first defined and used to describe the attack characteristics at diverse stages of s leak monitoring, hidden traffic overlay, and transaction identity forgery. The goal is to clarify how an Erebus attack affects the routing transfer and traffic state on the blockchain network layer. Consequently, detecting objects is expected to become more relevant and sensitive. A two-stage feature selection algorithm was designed based on ReliefF and weighted maximum relevance minimum redundancy (ReliefF_WMRmR) to alleviate the overfitting of the training model caused by redundant information and noise in multiple source features of the routing status and traffic behavior. The ReliefF algorithm was introduced to select strong correlations and highly informative features of the labeled data. According to WMRmR, a feature selection framework was defined to eliminate weakly correlated features, eliminate redundant information, and reduce the detection overhead of the model. A multimodal deep learning model was constructed based on the multilayer perceptron (MLP) to settle the high false alarm rates incurred by multisource data. Using this model, isolated inputs and deep learning were conducted on the selected routing status and traffic behavior. Redundant intermodal information was removed because of the complementarity of the multimodal network, which was followed by feature fusion and output feature representation to boost classification detection precision. The experimental results demonstrate that the proposed method can detect features, such as traffic data, at key link nodes and route messages in a real blockchain network environment. Additionally, the model can detect Erebus attacks effectively. This study provides novelty to the existing Erebus attack detection by increasing the accuracy detection by 1.05%, the recall rate by 2.01%, and the F1-score by 2.43%.     

An Intrusion Detection Algorithm Based on Feature Graph

With the development of Information technology and the popularization of Internet, whenever and wherever possible, people can connect to the Internet optionally. Meanwhile, the security of network traffic is threatened by various of online malicious behaviors. The aim of an intrusion detection system (IDS) is to detect the network behaviors which are diverse and malicious. Since a conventional firewall cannot detect most of the malicious behaviors, such as malicious network traffic or computer abuse, some advanced learning methods are introduced and integrated with intrusion detection approaches in order to improve the performance of detection approaches. However, there are very few related studies focusing on both the effective detection for attacks and the representation for malicious behaviors with graph. In this paper, a novel intrusion detection approach IDBFG (Intrusion Detection Based on Feature Graph) is proposed which first filters normal connections with grid partitions, and then records the patterns of various attacks with a novel graph structure, and the behaviors in accordance with the patterns in graph are detected as intrusion behaviors. The experimental results on KDD-Cup 99 dataset show that IDBFG performs better than SVM (Supprot Vector Machines) and Decision Tree which are trained and tested in original feature space in terms of detection rates, false alarm rates and run time.     

基于GADF-CNN的滚动轴承故障诊断方法

针对一维信号作为卷积神经网络输入时无法充分利用数据间的相关信息的问题,提出GADF-CNN的轴承故障诊断模型。利用格拉姆角差域(GADF)对采集到的振动信号进行编码,可以很容易地进行角度透视,从而识别出不同时间间隔内的时间相关性并生产相应特征图,之后将其输入卷积神经网络(CNN)自适应的完成滚动轴承故障特征的提取与分类。为了验证模型性能,采用凯斯西储大学轴承数据集进行轴承故障诊断分析,同时引入常见神经网络作为对比,检验不同模型的分类性能。结果表明,相较于其他图像编码方式与神经网络,该模型在载荷变化以及噪声污染时,仍保持了良好的诊断性能。     

基于深度置信网络与信息融合的齿轮故障诊断方法

针对齿轮在复杂运行工况下故障特征提取困难,传统故障诊断方法的识别精度易受人工提取特征的影响,以及单传感器获取信息不全面等问题,提出基于深度置信网络(DBN)与信息融合的齿轮故障诊断方法。通过多传感器信息融合技术对每个传感器采集的振动信号进行数据层融合;利用DBN进行自适应特征提取从而实现故障分类。为了避免因人为选择DBN结构参数,导致模型识别精度下降的问题,利用改进的混合蛙跳算法(ISFLA)对DBN结构参数进行优化。试验表明,与BP神经网络、未经优化的DBN以及单传感器故障诊断相比,该研究提出的信息融合及优化方法具有更高的故障识别精度。     

Monitoring a labeled degree-corrected stochastic block model

The stochastic block model (SBM) is a random graph model that focuses on partitioning the nodes into blocks or communities. A degree-corrected stochastic block model (DCSBM) considers degree heterogeneity within nodes. Investigation of the type of edge label can be useful for studying networks. We have proposed a labeled degree-corrected stochastic block model (LDCSBM), added the probability of the occurrence of each edge label, and monitored the behavior of this network. The LDCSBM is a dynamic network that varies over time; thus, we applied the monitoring process to both the US Senate voting network and simulated networks by defining structural changes. We used the Shewhart control chart for detecting changes and studied the effect of Phase I parameter estimation on Phase II performance. The efficiency of the model for surveillance has been evaluated using the average run length for estimated parameters.     

混凝土结构缺陷的融合识别研究

混凝土结构缺陷的无损检测是一项非常困难的工作,尤其是小尺寸或浅层的缺陷检测非常困难,由于仅用一种检测方法往往难以给出令人信服的结论,因此,本文同时采用超声检测和脉冲回波检测两种方法对不同大小的剥离和空洞缺陷进行了探测,并用小波分析方法对这两种信号进行了特征抽取,以第三阶尺度上的极大模作为信号的特征微量;随后,用一个多层前馈神经网络进行了单种检测方法的软决策,定义并计算了以该软决策为基础的概率分配函数;最后,用证据理论方法进行了两种检测方法的决策级融合识别,分类试验结果表明融合识别确实好于单一方法的识别。     

基于多尺度稠密卷积网络的单图像超分辨率重建

目的针对已有网络对于卷积特征图利用率低下,从而导致高倍数图像重建质量不高的情况,提出一种多尺度稠密卷积网络(SRMD)。方法对SRDenseNet的稠密连接模块进行改进,去除批规范化层,参考已有网络,设计多尺度特征提取层和1×1的信息整合层,从而构成多尺度稠密卷积模块。SRMD通过一个多尺度特征提取层堆叠64个底层特征图,再由8个多尺度稠密卷积模块经过稠密连接堆叠1024个特征图,最后通过信息整合和子像素卷积模块输出超分辨率重建图像。结果在Set5,Set14,B100和U100数据集上进行测试,SRMD重建图像的峰值信噪比分别为30.1570,26.9952,25.7860, 23.4821 dB,结构相似性分别为0.8813,0.7758,0.7243,0.7452。结论与已有网络相比,SRMD与DRCN,VDSR表现相当,优于SRDenseNet和BiCubic方法。     

基于卷积神经网络和Transformer网络的鸟声识别

针对传统鸟声识别算法中特征提取方式单一、分类识别准确率低等问题,提出一种结合卷积神经网络和Transformer网络的鸟声识别方法。该方法综合考虑网络局部特征学习和全局上下文依赖性构造,从原始鸟声音频信号中提取短时傅里叶变换(Short Time Fourier Transform,STFT)语谱图特征,将其输入到卷积神经网络(ConvolutionalNeural Network,CNN)中提取局部频谱特征信息,同时提取鸟声信号的对数梅尔特征及一阶差分、二阶差分特征用于合成梅尔频率倒谱系数(Mel Frequency Cepstrum Coefficient,MFCC)混合特征向量,将其输入到Transformer网络中获取全局序列特征信息,最后融合所提取的特征可得到更丰富的鸟声特征参数,通过Softmax分类器得到鸟声识别结果。在Birdsdata和xeno-canto鸟声数据集上进行实验,平均识别准确率分别达到了97.81%和89.47%。实验结果表明该方法相较于其他现有的鸟声识别模型具有更高的识别准确率。     

基于GST-ECNN的运动想象脑电信号识别方法

在对脑电信号的解码研究中,存在着现有时频分析方法对高频信号处理能力有限,多通道信号信息冗余,常用卷积神经网络分类器ReLU激活函数受学习速率的影响较大,对不同层采用相同的正则化很难获得满意结果等问题。为此,提出了一种基于广义S变换特征提取和增强卷积神经网络分类相结合的方法,同时提出一种结合Relief算法和向前选择搜索策略的包裹式方法进行通道选择。结果表明,提出的方法利用较少的信号通道,具有更强的特征提取和分类的能力,在第Ⅳ届BCI的数据集I上取得最高98.44±1.5%的分类准确率,高于其他现有算法。该方法良好的分类性能不仅减少了计算消耗,也有效提高了分类准确率,对脑电信号特征提取和分类具有一定的参考意义。     

A Network Traffic Classification Model Based on Metric Learning

Attacks on websites and network servers are among the most critical threats in network security. Network behavior identification is one of the most effective ways to identify malicious network intrusions. Analyzing abnormal network traffic patterns and traffic classification based on labeled network traffic data are among the most effective approaches for network behavior identification. Traditional methods for network traffic classification utilize algorithms such as Naive Bayes, Decision Tree and XGBoost. However, network traffic classification, which is required for network behavior identification, generally suffers from the problem of low accuracy even with the recently proposed deep learning models. To improve network traffic classification accuracy thus improving network intrusion detection rate, this paper proposes a new network traffic classification model, called ArcMargin, which incorporates metric learning into a convolutional neural network (CNN) to make the CNN model more discriminative. ArcMargin maps network traffic samples from the same category more closely while samples from different categories are mapped as far apart as possible. The metric learning regularization feature is called additive angular margin loss, and it is embedded in the object function of traditional CNN models. The proposed ArcMargin model is validated with three datasets and is compared with several other related algorithms. According to a set of classification indicators, the ArcMargin model is proofed to have better performances in both network traffic classification tasks and open-set tasks. Moreover, in open-set tasks, the ArcMargin model can cluster unknown data classes that do not exist in the previous training dataset.     

Using Object Detection Network for Malware Detection and Identification in Network Traffic Packets

In recent years, the number of exposed vulnerabilities has grown rapidly and more and more attacks occurred to intrude on the target computers using these vulnerabilities such as different malware. Malware detection has attracted more attention and still faces severe challenges. As malware detection based traditional machine learning relies on exports’ experience to design efficient features to distinguish different malware, it causes bottleneck on feature engineer and is also time-consuming to find efficient features. Due to its promising ability in automatically proposing and selecting significant features, deep learning has gradually become a research hotspot. In this paper, aiming to detect the malicious payload and identify their categories with high accuracy, we proposed a packet-based malicious payload detection and identification algorithm based on object detection deep learning network. A dataset of malicious payload on code execution vulnerability has been constructed under the Metasploit framework and used to evaluate the performance of the proposed malware detection and identification algorithm. The experimental results demonstrated that the proposed object detection network can efficiently find and identify malicious payloads with high accuracy.     

人机接触运动交互中人体舒适感知的测量方法

目的 完善人机接触运动交互中用户的舒适度测量方法,提高机器人作业的智能感知能力。方法 基于机器人按摩作业平台,收集测试者人机接触中皮肤所受的力与变形等作业数据,以及主观触觉感知评价。建立以疼痛感、牵扯感和灼热感为基本维度的三维舒适度感知空间,通过反向传播神经网络建立客观测量数据与人体触觉感知的映射关系;采用支持向量机算法对人体触觉感知进行舒适、不舒适可作业和不舒适需停机分类。结果 实验结果显示,触觉感知预测模型和状态分类模型,能有效实现从作业特征信息到触觉感知再到人机作业状态的整体辨识。结论 该方法对人机交互过程中的作业特征和人体主观触觉感知进行联系和模式识别,为机器人安全、舒适的作业规划和在线控制提供了有效依据。     

Toward Robust Classifiers for PDF Malware Detection

Malicious Portable Document Format (PDF) files represent one of the largest threats in the computer security space. Significant research has been done using handwritten signatures and machine learning based on detection via manual feature extraction. These approaches are time consuming, require substantial prior knowledge, and the list of features must be updated with each newly discovered vulnerability individually. In this study, we propose two models for PDF malware detection. The first model is a convolutional neural network (CNN) integrated into a standard deviation based regularization model to detect malicious PDF documents. The second model is a support vector machine (SVM) based ensemble model with three different kernels. The two models were trained and tested on two different datasets. The experimental results show that the accuracy of both models is approximately 100%, and the robustness against evasive samples is excellent. Further, the robustness of the models was evaluated with malicious PDF documents generated using Mimicus. Both models can distinguish the different vulnerabilities exploited in malicious files and achieve excellent performance in terms of generalization ability, accuracy, and robustness.     

基于灵敏度预分类的BP神经网络故障诊断方法

故障特征信息的获取和处理对电路故障的可靠分类和准确诊断有很大的影响．在电路故障诊断时,对于不同的故障模式,存在信息混叠的现象,需要解决特征信息的有效提取和故障的可靠分类等问题．为此,本文提出了一种结合灵敏度特性分析的BP神经网络故障诊断方法．基本思想是通过灵敏度的计算,对电路故障样本作预分类,再根据电路灵敏度的计算结果分别提取相应特征信息,以此构造故障样本特征集,然后作为BP神经网络的输入对网络训练,并进行故障诊断．对滤波器的仿真结果表明,该方法能分类不同的元件故障,且对模拟电路故障诊断的平均正确率优于传统方法．     

Malicious Traffic Detection in IoT and Local Networks Using Stacked Ensemble Classifier

Malicious traffic detection over the internet is one of the challenging areas for researchers to protect network infrastructures from any malicious activity. Several shortcomings of a network system can be leveraged by an attacker to get unauthorized access through malicious traffic. Safeguard from such attacks requires an efficient automatic system that can detect malicious traffic timely and avoid system damage. Currently, many automated systems can detect malicious activity, however, the efficacy and accuracy need further improvement to detect malicious traffic from multi-domain systems. The present study focuses on the detection of malicious traffic with high accuracy using machine learning techniques. The proposed approach used two datasets UNSW-NB15 and IoTID20 which contain the data for IoT-based traffic and local network traffic, respectively. Both datasets were combined to increase the capability of the proposed approach in detecting malicious traffic from local and IoT networks, with high accuracy. Horizontally merging both datasets requires an equal number of features which was achieved by reducing feature count to 30 for each dataset by leveraging principal component analysis (PCA). The proposed model incorporates stacked ensemble model extra boosting forest (EBF) which is a combination of tree-based models such as extra tree classifier, gradient boosting classifier, and random forest using a stacked ensemble approach. Empirical results show that EBF performed significantly better and achieved the highest accuracy score of 0.985 and 0.984 on the multi-domain dataset for two and four classes, respectively.