Network security situation assessment and prediction method based on multimodal transformation in edge computing

The continuous development of information technology has also promoted the progress of the Internet. More people are joining the Internet. The amount of data stored in the network is also increasing, including some important information, which leads to criminals launching attacks on network security. In order to solve the large error in network security situation assessment and poor progress in network security prediction, the study uses spectrum clustering analysis to evaluate the network security situation. Then genetic algorithm, grey wolf optimization algorithm and support vector machine fusion algorithm are used to predict the Network Security Service (NSS). The genetic algorithm is used to optimize the global search ability of the gray wolf optimization algorithm and the parameters of the support vector machine are optimized to evaluate and predict the NSS. The results showed that the maximum error of the proposed model was 0.4112, and the maximum error was 0.5896. The absolute percentage error of this algorithm was 0.0270, while the other algorithms were 0.0745 and 0.0952, respectively. The proposed model has lower errors and time consumption in training and simulation testing compared with other current methods. The network situation assessment and prediction method proposed in the study can effectively improve network security services, ensure the personal information security, and enhance the security of the Internet.

Background: Radiotherapy (RT) is a cornerstone of multimodal treatment for rectal cancer (RC); yet, substantial interindividual variability in treatment response persists. Deep learning (DL)–based radiomics offers potential for pre-RT response prediction to support personalized decision-making.Objective: This study aimed to develop and compare multiple DL radiomics models for predicting RT response in RC, with emphasis on the performance and clinical utility of Transformer architectures.Methods: In this single-center retrospective study, 2000 pathologically confirmed patients with RC who received standard RT were included. Pretreatment computed tomographic and dynamic contrast-enhanced magnetic resonance images and clinical variables were collected. Treatment response was categorized according to RECIST (Response Evaluation Criteria in Solid Tumors) version 1.1 as good (complete or partial response) or poor (stable or progressive disease). The primary analysis used magnetic resonance imaging (MRI)–only input; computed tomography (CT) was used for registration and quality control and evaluated in a late-fusion CT + MRI sensitivity analysis. Data were randomly split into training, validation, and test sets (8:1:1), with 5-fold cross-validation within the training set. Test set tumor masks were manually delineated, whereas a U-Net assisted segmentation was performed only within training to prevent data leakage. Convolutional neural network, graph convolutional network, and Transformer classifiers were compared. Class imbalance (approximately 65% vs 35%) was addressed using class weighting. Performance was evaluated using area under the receiver operating characteristic curve (AUROC) and accuracy with 95% CIs obtained by bootstrapping. AUROC differences were assessed using the DeLong test. Clinical usefulness was evaluated using decision curve analysis. Segmentation performance was quantified by Dice coefficient and intersection over union. Model interpretability was assessed using Gradient-Weighted Class Activation Mapping.Results: In the MRI-only primary analysis, the Transformer achieved the best performance on the independent test set, with accuracy of 87.0% (95% CI 84.2%-89.5%) and AUROC of 0.921 (95% CI 0.901-0.945), significantly outperforming the convolutional neural network (AUROC 0.881; P=.02) and graph convolutional network (AUROC 0.894; P=.041). Sensitivity and specificity were 89.2% and 82.9%, respectively. Decision curve analysis demonstrated higher net benefit across threshold probabilities of 0.3-0.7. U-Net segmentation achieved a mean Dice coefficient of 0.892 and intersection over union of 0.814. In sensitivity analysis, CT + MRI late fusion yielded a comparable AUROC to MRI only (0.926 vs 0.921; P=.36), with modest incremental net benefit at higher thresholds.Conclusions: In this large pre-RT imaging cohort, an MRI-driven Transformer-based DL radiomics model outperformed conventional architectures in predicting RT response in RC and demonstrated superior clinical net benefit. Late fusion of CT and MRI did not significantly improve overall discrimination but may provide incremental benefit in specific decision contexts. Multicenter external validation is warranted.

Breast cancer is the most common cancer worldwide, and magnetic resonance imaging (MRI) constitutes a very sensitive technique for invasive cancer detection. When reviewing breast MRI examination, clinical radiologists rely on multimodal information, composed of imaging data but also information not present in the images such as clinical information. Most machine learning (ML) approaches are not well suited for multimodal data. However, attention-based architectures, such as Transformers, are flexible and therefore good candidates for integrating multimodal data. The aim of this study was to develop and evaluate a novel multimodal deep learning (DL) model combining ultrafast dynamic contrast-enhanced (UF-DCE) MRI images, lesion characteristics and clinical information for breast lesion classification. From 2019 to 2023, UF-DCE breast images and radiology reports of 240 patients were retrospectively collected from a single clinical center and annotated. Imaging data were constituted of volumes of interest (VOI) extracted around segmented lesions. Non-imaging data were constituted of both clinical (categorical) and geometrical (scalar) data. Clinical data were extracted from annotated reports and were associated to their corresponding lesions. We compared the diagnostic performances of traditional ML methods for non-imaging data, an image model based on the DL architecture, and a novel Transformer-based architecture, the Multimodal Sieve Transformer with Vision Transformer encoder (MMST-V). The final dataset included 987 lesions (280 benign, 121 malignant lesions, and 586 benign lymph nodes) and 1081 reports. For classification with non-imaging data, scalar data had a greater influence on performances of lesion classification (Area under the receiver operating characteristic curve (AUROC)=0.875±0.042) than categorical data (AUROC=0.680±0.060). MMST-V achieved better performances (AUROC=0.928±0.027) than classification based on non-imaging data (AUROC=0.900±0.045), and imaging data only (AUROC=0.863±0.025). The proposed MMST-V is an adaptative approach that can consider redundant information provided by multimodal information. It demonstrated better performances than unimodal methods. Results highlight that the combination of clinical patient data and detailed lesion information as additional clinical knowledge enhances the diagnostic performances of UF-DCE breast MRI.

Recently, the network security has become a severe issue. Obtaining the current network security status and predicting the future network security status may contribute to the decisions given by the decision makers. Under the large-scale and complex network environment, the traditional network management method cannot completely meet the demand of network situation awareness, and the network situation awareness based on information fusion has become the development direction of the future. Because network situation assessment is the core of network situation awareness, this paper introduces fuzzy rough set theory based on technologies of network security situation assessment, overcoming the defect that rough set method has to discretize while lost accuracy. By combining the advantages of the fuzzy set and rough set in dealing with uncertainty and vagueness of the data, this paper establishes a network security situation assessment model based on fuzzy rough set. The experiment shows that the model can provide high accuracy, and can provide a more accurate and reasonable evaluation results for the network security situation assessment.

Network security situation assessment (NSSA) is an important and effective active defense technology in the field of network security situation awareness. By analyzing the historical network security situation awareness data, NSSA can evaluate the network security threat and analyze the network attack stage, thus fully grasping the overall network security situation. With the rapid development of 5G, cloud computing, and Internet of things, the network environment is increasingly complex, resulting in diversity and randomness of network threats, which directly determine the accuracy and the universality of NSSA methods. Meanwhile, the indicator data is characterized by large scale and heterogeneity, which seriously affect the efficiency of the NSSA methods. In this paper, we design a new NSSA method based on the autoencoder (AE) and parsimonious memory unit (PMU). In our novel method, we first utilize an AE‐based data dimensionality reduction method to process the original indicator data, thus effectively removing the redundant part of the indicator data. Subsequently, we adopt a PMU deep neural network to achieve accurate and efficient NSSA. The experimental results demonstrate that the accuracy and efficiency of our novel method are both greatly improved.

To solve the problems that existing network security situation assessment (NSSA) methods are difficult to extract features and have poor timeliness, an NSSA method with network attack behavior classification (NABC) is proposed. First, an NABC model is designed. The model combines features and advantages of a parallel feature extraction network (PFEN), a bidirectional gate recurrent unit (BiGRU), and the attention mechanism (ATT). The PFEN module is composed of parallel sparse autoencoders which extract key data from different network attack behaviors. The BiGRU module gets the time-series relationship from the state of three different time periods, finds potential representation rules from network attack behaviors. The ATT module pays more attention to the network traffic key information and improves the NABC accuracy. Second, the NABC detects and classifies attacks from network behaviors, the occurrence number of each attack behavior, and the error probability matrix are counted. Finally, the occurrence number of each attack behavior is corrected according to the error probability matrix, and the network security situation value is calculated through combining the severity factor of each attack behavior. The experimental results show that the precision and recall of the NABC model are improved by 5.28% and 5.65%, respectively, compared with the conventional method. The comparison experiment with the classical situation assessment method also proves that the proposed method can assess the overall situation of network security more effectively and comprehensively.

Graph neural networks could compare the difference among all samples (nodes in graph) and transmit the interrelationship among them to obtain a global landscape. Compared with radiomics and clinical feature‐based machine learning methods, whether a graph convolutional neural network (GCNN) based on radiomics and clinical features improve the performance in distinguishing benign and malignant pulmonary nodules is not well studied. We propose an approach based on multimodal GCNNs that integrates patients’ lung computed tomography images with clinical information to differentiate between benign and malignant pulmonary nodules. Leveraging large‐scale and multisource data from multiple hospitals (i.e., 6033/290/524 patients for three hospitals respectively) enhances the diversity of features. Accuracy, sensitivity, specificity, precision, and area under the receiver operating characteristic curve (AUROC) are used to evaluate the performance. We achieved the average accuracy/sensitivity/specificity/AUROC of 0.8612/0.9425/0.6786/0.9025 for the main dataset via the novel GCNN proposed, respectively, maintaining the robustness of the deep learning procedures. Especially for the external testing dataset (hospital 2/hospital 3), the specificity is much higher than comparison methods (0.6250–0.6731 vs. 0.2569–0.2788). The graph neural network‐based deep learning method holds the potential to assist clinicians, aiding in treatment planning, patient management, follow‐up strategies, resource optimization, and overall healthcare decision‐making.

Network security refers to the ability to protect computer networks from unauthorized access, destruction, theft or damage. With the popularization of the Internet and the acceleration of information development, network security has become a very important issue. Network security situational awareness is a new network security protection technology that evaluates the security status of the current network operating environment and predicts the security change trend of the network operating environment in the future period based on the relevant elements in the current network operating environment, with the ultimate purpose of ensuring network security. At present, researches on network security situation approach mainly focus on improving the accuracy of logical judgment, which is difficult to achieve a balance between logical processing and spatio-temporal complexity. In addition, there are some problems such as high complexity and lack of objectivity of the implementation model. Based on this, this paper designs a new network security situation assessment method based on D-S evidence theory. Firstly, aiming at the deficiency of objectivity in traditional single D-S evidence theory, Deep Neural Networks (DNN) is proposed to obtain OP to reduce its subjective dependence. Secondly, in order to solve the problems of low accuracy and low time efficiency when DNN processes massive data, SAE is used to reduce the dimensionality of massive high-dimensional data, and D-DNNsafe network security situation assessment model is constructed.

In order to strengthen the power network structure of our country, reduce the frequency of electrical equipment and lines, so that the power network can operate normally and smoothly. Based on the above reasons, the flow feature association mining algorithm is introduced into the research of power network security situation assessment to evaluate and predict the safe running state of power network. Firstly, this paper explains the basic algorithm of stream feature mining, including time correlation, IP (Ingress Protection) correlation, type correlation and super warning correlation. Secondly, it is necessary to determine the weight of network security evaluation index and the calculation method of evaluation, and carry out the relevant case analysis. Finally, the network security situation assessment model based on the flow feature association mining algorithm was compared with the BP(Back Propagation) neural network and RBF(Radial Basis Function) neural network evaluation model, and the comparison results showed that Flow feature association mining algorithm can improve the feasibility and accuracy of power network situation assessment. In conclusion, the study of this paper can provide a new research approach for the construction of network security situation assessment.

Network security situation awareness is a new type of network security technology. It evaluates the network security situation in real time from a macro perspective. Also it can predict the trend of the development of the network security situation, providing a basis for the decision analysis of administrators. It is difficult to obtain complete and accurate information in network security situation assessment by using evidential network. So we introduce an evidential network based on Bayesian network to solve that problem. Firstly, transform the parent node information and inference rules into plausibility function so as to be compatible with imperfect and inaccurate information. Secondly, we use the full probability formula of Bayesian network as reference to make similar reasoning under the framework of evidence theory. Then transform the inference result to BPA form by using the minimum specificity algorithm, and obtain the final result by projection. Finally, an example of network security situation assessment is given to illustrate the rationality and effectiveness of the method.KeywordsNetwork security situation assessmentEvidence theoryBayesian networkEvidential network

With the expansion of power system networks, network security issues have become increasingly prominent. The wide coverage and complex nodes of power grids make them highly vulnerable to network attacks, which can have severe consequences. Traditional security measures for power system networks are unable to efficiently detect and warn against various types of network attacks. Therefore, this paper proposes a network security situation assessment method based on CBAM-EfficientNet for anomaly detection. Firstly, we employ the CBAM-EfficientNet model for the detection and classification of anomalies, adjusting the count of attack types through a false positive probability matrix. Secondly, we calculate the network security situation value by integrating the severity and impact of the attacks. Finally, we evaluate the network security situation, which aids management personnel in gaining a comprehensive understanding of the security status. Experimental results show that our CBAM-EfficientNet model outperforms mainstream deep learning models, such as Xception, ResNet, LSTM, and GRU, by improving accuracy and F1 score by approximately 5%. The comparative experiments in network security situation assessment also prove the effectiveness and comprehensiveness of this method in evaluating the overall network security situation.

Network security situational assessment, the core task of network security situational awareness, can obtain security situation by comprehensively analyzing various factors that affect network status. Thus, network security situational assessment can provide accurate security state evaluation and security trend prediction for users. Although plenty of network security situational assessment methods have been proposed, there are still many problems to solve. First, because of high dimensionality of input data, computational complexity in model construction could be very high. Moreover, most of the existing schemes trade computational overhead for accuracy. Second, due to the lack of centralized standard, the weights of indicators are usually determined empirically or by subjective opinions of domain expert. To solve the above problems, we propose a novel network security situation assessment method based on stack autoencoding network and back propagation neural network. In stack autoencoding network and back propagation neural network, to reduce the data storage overhead and improve computational efficiency, we use stack autoencoding network to reduce the dimensions of the indicator data. And the low-dimensional data output by hidden layer of stack autoencoding network will be the input data of the error back propagation neural network. Then, the back propagation neural network algorithm is adopted to perform network security situation assessment. Finally, extensive experiments are conducted to verify the effectiveness of the proposed method.

Precipitation nowcasting, aiming to predict the rainfall intensity in the near future (usually 0-2h) [1], is crucial for urban planning, flood monitoring, agriculture management, and so on. Numerical weather modeling (NWP) takes a variety of data sources as the input of complex computer models that use mathematical equations to simulate the behavior of the atmosphere. Limited by the time needed for model spin-up, the performance in the short near future is not satisfactory. Deep learning (DL)-based method fills in the gap by treating nowcasting as a video prediction problem. The Convolutional LSTM [2] extracts spatial information when dealing with temporal series. The Generated Adversarial Network (GAN)-based [3] method shows potential in simulating the realisticness of the precipitation field. However, training such a model is very time-consuming and data-demanding [3] [4]. Different from natural images, the precipitation field to be estimated usually has a larger spatial size. Moreover, the convolutional layers tend to oversmooth the output and eliminate the small patterns that are important for the meteorologists to make the decision. Thus, we proposed a two-stage framework: the first one is to train an RNN-based model on the coarse field. The second is to downscale and style transfer from the coarse field to high-resolution precipitation maps based on GAN and Graph Convolutional Network (GCN). The coarse prediction will act as a constraint to the finer scale output and allows re-assignment of the spatial distribution of intensities. Such probabilistic output prevents the overestimation of the intensity. RNN is good at capturing long-range characteristics, and GCN [5] can extract local and neighborhood information, thus these two channels are naturally complementary to improve both local patterns and global accuracy scores. The GAN is used to make final output similar to real precipitation maps such as radar scans. To train the model, we downloaded the 2006-2016 ERA5 total precipitation at 0.25-degree spatial resolution and the DWD radar map [6] at 1km spatial resolution. We expect our model can capture the overall coverage of rainfall events and depict the spatial details. More importantly, this alleviates the data shortage problem, i.e., high-resolution precipitation nowcasting at places without ground-based radar stations can be acquired. [1] Shi, Xingjian, et al. "Deep learning for precipitation nowcasting: A benchmark and a new model." Advances in neural information processing systems 30 (2017).[2]Shi, Xingjian, et al. "Convolutional LSTM network: A machine learning approach for precipitation nowcasting." Advances in neural information processing systems 28 (2015).[3] Ravuri, Suman, et al. "Skilful precipitation nowcasting using deep generative models of radar." Nature 597.7878 (2021): 672-677.[4] Sønderby, Casper Kaae, et al. "Metnet: A neural weather model for precipitation forecasting." arXiv preprint arXiv:2003.12140 (2020).[5] Shi, Yilei, Qingyu Li, and Xiao Xiang Zhu. "Building segmentation through a gated graph convolutional neural network with deep structured feature embedding." ISPRS Journal of Photogrammetry and Remote Sensing 159 (2020): 184-197.[6] Ayzel, Georgy, Tobias Scheffer, and Maik Heistermann. "RainNet v1. 0: a convolutional neural network for radar-based precipitation nowcasting." Geoscientific Model Development 13.6 (2020): 2631-2644.

With the rapid development of the information society, security threats in cyberspace are increasing day by day, posing severe challenges to national infrastructure, commercial operations, and even personal privacy. At present, the research of network security situational awareness and risk assessment model is faced with critical problems, such as significant demand for prior knowledge, high complexity of inference algorithm, insufficient dynamic adaptability, and inaccurate identification of risk categories. In view of this, this study proposes a new network security situational awareness and risk assessment model based on the Bayesian network, aiming to achieve timely early warning and accurate prediction of network threats through probability statistics methods. By comprehensively considering various heterogeneous data sources such as network traffic anomalies, system log anomalies, and external threat intelligence, we built a sizeable Bayesian network covering thousands of nodes and hundreds of thousands of edges to describe the occurrence mechanism and evolution path of network security incidents. Empirical research shows that the optimized model has an accuracy rate of 92%, a recall rate of 89%, and an F1 score of 90.5% on the test dataset, which is significantly better than the existing rule-based and machine-learning methods, especially when dealing with low-frequency threats with apparent long-tail effects, showing more robust adaptability and prediction accuracy. By dynamically monitoring the changing trend of network activities, we can identify potential risk points in advance, help take proactive protective measures before security threats occur, and effectively reduce economic losses caused by network intrusions. This study not only provides a brand-new theoretical framework and technical means for network security situational awareness and risk assessment but also opens up broad prospects for subsequent research directions and application scenarios.

ObjectiveExcessive prescription of antibiotics is amongst the principal drivers of antibiotic resistance, which is considered a surging threat to global health. The most frequent resistant pathogens are usually linked with...