Network-based Network Research Articles

CNN-based Network Intrusion Detection and Classification Model for Cyber-Attacks

A Convolution Neural Network (CNN)-based Network Intrusion Detection Model for Cyber-attacks is of great value in identifying and classifying attacks on any network. The Knowledge Discovery in Database Cup '99 dataset containing approximately 4,900,000 single connection vectors was divided into two phases; 75% of the total dataset was used during the learning process of the machine learning technique, while 25% was used on a fully trained model to validate and evaluate its performance. The model's performance indicated that it can detect and classify different classes of attacks with an accuracy of 98% with 20 epochs at a 0.001 learning rate using machine learning. The model loss for the training and validation was 7.48% and 7.98%, respectively, over 20 epochs, which implies that the model performed better on the training dataset. This study demonstrated that the convolutional Neural network-based Network Intrusion Detection and classification model shows high detection and low false negative rates. The CNN model offers a high detection rate and fidelity to unknown attacks, i.e., it can differentiate between already-seen attacks and new zero-day attacks. At the end of the experiment, the proposed approach is suitable in modeling the network IDS for detecting intrusion attacks on computer networks thereby enabling a secured environment for the proper functioning of the system

A Neural Network-Based Network Selection for QUIC to Enrich Gaming in NextGen Wireless Network

A Neural Network-Based Network Selection for QUIC to Enrich Gaming in NextGen Wireless Network

Optimalisasi Dua Layanan Jaringan Internet Menggunakan Teknik Load Balancing dengan Metode Peer Connection Classifier (PCC) (Studi Kasus: Jaringan Internet Desa Banyuanyar Boyolali)

Load Balancing is a technique for distributing data traffic so that the workload is distributed evenly on 2 or more networks to maximize resource usage and improve performance. There are two ISPs for internet access services to the Banyuanyar Village Office, Boyolali, both of which have different bandwidths, namely from PT. Telkom has 100Mbps bandwidth and Kominfo has 50Mbps bandwidth. So here there will be dense network traffic if the internet network is not optimized.. In this case, it can be optimized using load balancing techniques on both ISPs by using the Peer Connection Classifier technique to group traffic and divide the load on both internet connection lines so as not to overload. So it can be concluded that by adding a proxy router device, configuring load balancing, and applying a hierarchical network-based network topology, it will optimize two internet network services at Banyuanyar Village Hall, Boyolali.

GraphNEI: A GNN-based network entity identification method for IP geolocation

Network entity geolocation technology is a technique for inferring geographic location through features such as network measurements or IP address searchable information, also known as IP geolocation. By obtaining the device type of the target can assist in target IP geolocation or IP landmark mining. Most traditional rule-based or learning-based methods identify network entities. However, due to the existence of firewalls and the limitations of feature detection, the features of individual targets are prone to be missing or spoofed, which can lead to a decrease in the accuracy of network entity identification (NEI). This paper propose a graph neural network-based network entity identification method, GraphNEI model, which embeds network entities into the graph structure and uses graph neural networks to improve the current problem of missing or spoofed features in order to improve the accuracy of current network entity identification. It mainly includes five steps: data processing, subgraph partition, weight calculation, node update and classification. First, the acquired dataset is subjected to feature extraction and anonymization; second, the target nodes are subjected to network topology graph construction and community division; third, the nodes’ self-attention, structural attention and similarity of neighbors are combined and used to calculate the nodes’ combined attention; fourth, node aggregation and update, and update the node representation based on the combined attention results; finally, the nodes are classified. We successfully identified 7 different network entities on the publicly collected dataset, and the identification accuracy of network entities is above 95.49%, improved 0.51%–10.42% compared to typical rule-based or learning-based methods. It effectively improves the effective NEI in the absence of network entity features. In addition, we conduct IP geolocation research based on NEI, and the experimental results show that the method is effective in reducing IP geolocation errors distance.

ResMIBCU-Net: an encoder-decoder network with residual blocks, modified inverted residual block, and bi-directional ConvLSTM for impacted tooth segmentation in panoramic X-ray images.

Impacted tooth is a common problem that can occur at any age, causing tooth decay, root resorption, and pain in the later stages. In recent years, major advances have been made in medical imaging segmentation using deep convolutional neural network-based networks. In this study, we report on the development of an artificial intelligence system for the automatic identification of impacted tooth from panoramic dental X-ray images. Among existing networks, in medical imaging segmentation, U-Net architectures are widely implemented. In this article, for dental X-ray image segmentation, blocks and convolutional block structures using inverted residual blocks are upgraded by taking advantage of U-Net's network capacity-intensive connections. At the same time, we propose a method for jumping connections in which bi-directional convolution long short-term memory is used instead of a simple connection. Assessment of the proposed artificial intelligence model performance was evaluated with accuracy, F1-score, intersection over union, and recall. In the proposed method, experimental results are obtained with 99.82% accuracy, 91.59% F1-score, 84.48% intersection over union, and 90.71% recall. Our findings show that our artificial intelligence system could help with future diagnostic support in clinical practice.

Bearing fault diagnosis using a speed-adaptive network based on vibro-speed data fusion and majority voting

Fault diagnosis of rolling bearings is key to maintain and repair modern rotating machinery. Rolling bearings are usually working in non-stationary conditions with time-varying loads and speeds. Existing diagnosis methods based on vibration signals only do not have the ability to adapt to rotational speed. And when the load changes, their accuracy rate will be obviously reduced. A method is put forward which fuses multi-modal sensor signals to fit speed information. Firstly, the features are extracted from raw vibration signals and instantaneous rotating speed signals, and fused by 1D-convolution neural network-based networks. Secondly, to improve the robustness of the model when the load changes, a majority voting mechanism is proposed in the diagnosis stage. Lastly, multiple variable speed samples of four bearings under three loads are obtained to evaluate the performance of the proposed method by analyzing the loss function, accuracy rate and F 1 score under different variable speed samples. It is empirically found that the proposed method achieves higher diagnostic accuracy and speed-adaptive ability than the algorithms based on vibration signal only. Moreover, a couple of ablation studies are also conducted to investigate the inner mechanism of the proposed speed-adaptive network.

Challenges of deep learning methods for COVID-19 detection using public datasets

Since the COVID-19 pandemic, several research studies have proposed Deep Learning (DL)-based automated COVID-19 detection, reporting high cross-validation accuracy when classifying COVID-19 patients from normal or other common Pneumonia. Although the reported outcomes are very high in most cases, these results were obtained without an independent test set from a separate data source(s). DL models are likely to overfit training data distribution when independent test sets are not utilized or are prone to learn dataset-specific artifacts rather than the actual disease characteristics and underlying pathology. This study aims to assess the promise of such DL methods and datasets by investigating the key challenges and issues by examining the compositions of the available public image datasets and designing different experimental setups. A convolutional neural network-based network, called CVR-Net (COVID-19 Recognition Network), has been proposed for conducting comprehensive experiments to validate our hypothesis. The presented end-to-end CVR-Net is a multi-scale-multi-encoder ensemble model that aggregates the outputs from two different encoders and their different scales to convey the final prediction probability. Three different classification tasks, such as 2-, 3-, 4-classes, are designed where the train–test datasets are from the single, multiple, and independent sources. The obtained binary classification accuracy is 99.8% for a single train–test data source, where the accuracies fall to 98.4% and 88.7% when multiple and independent train–test data sources are utilized. Similar outcomes are noticed in multi-class categorization tasks for single, multiple, and independent data sources, highlighting the challenges in developing DL models with the existing public datasets without an independent test set from a separate dataset. Such a result concludes a requirement for a better-designed dataset for developing DL tools applicable in actual clinical settings. The dataset should have an independent test set; for a single machine or hospital source, have a more balanced set of images for all the prediction classes; and have a balanced dataset from several hospitals and demography. Our source codes and model are publicly available11https://github.com/kamruleee51/CVR-Net. for the research community for further improvements.

Combined RBFN based MPPT and d-axis stator current control for permanent magnet synchronous generators

This paper proposes a new radial basis function neural network maximum power point tracking controller based on a differential evolution algorithm for machine side converter of permanent magnet synchronous generator wind turbine under variable wind speed. Direct axis stator current control methods of permanent magnet synchronous machine are reviewed shortly. A combined radial basis function neural network-based network maximum power point tracking method and d axis stator current control techniques including zero d axis stator current, unity power factor, and constant stator flux-linkage have been implemented to control the machine side converter of permanent magnet synchronous generator wind turbine. The dynamic performance of the proposed approach is assessed under different operating conditions through a simulation model based on MATLAB. It has been seen that the radial basis function neural network controller can not only track well the maximum power point but also can be reduced costly.

Event Prediction Technology Based on Graph Neural Network

With the rapid development of the mobile Internet and the continuous expansion of network scale, the network security situation is becoming increasingly severe, and the endless network security threats have put forward higher requirements for network security performance. Based on the above background, the purpose of this paper is to explore the event prediction technology based on graph neural network. Due to the slow convergence of the network event prediction and evaluation model, the untimely risk assessment and inaccurate safety prediction caused by the incomplete parameter setting of the prediction model have become prominent problems in this field. This paper proposes an event prediction technology based on graph neural network. This method first uses genetic algorithms to optimize the weights in the training process of the graph neural network, which overcomes the blindness of initial weight selection and improves the training efficiency of the graph neural network; the KDDCup99 data set is used to conduct experiments on the above two methods respectively. Verification and analysis. The simulation and comparison experiments respectively verify that the neural network-based network security situation assessment and prediction method proposed in this paper can realize the assessment and prediction of the network situation more efficiently and accurately.

Predicting ready-made garment dressing fit for individuals based on highly reliable examples

Predicting the dressing fit of a ready-made garment for different individuals is a challenging problem in online garment sales. At present, the main solution involves physics-based virtual garment simulation, which has limitations in terms of a lack of sufficient reality and high computational costs. In this study, we developed a novel example-based method to guarantee the high reality and efficiency of garment dressing fit prediction using two methods. First, highly reliable examples were captured with the assistance of a robotic mannequin, thereby ensuring the authenticity of the sample data. Second, a mapping was established between the body-garment ease allowance and garment deformations using a convolutional neural network-based network in order to address the problem of dressing fit prediction for ready-made garments on different individuals. This method can also be extended to predicting the dressing fit for ready-made garments with similar styles. Experiments showed our virtual clothes try-on system obtained acceptable precision with a good sense of reality, and it can potentially be used in many applications such as three-dimensional garment design and online clothes retail.

WOA-NN: a decision algorithm for vertical handover in heterogeneous networks

The heterogeneous network of a 4th generation not always support better communication and mobility between the Wireless Access Networks. Hence, the vertical handoff is highly necessitated. This paper establishes vertical handover, which is context-aware in a heterogeneous environment with WiMax and WiFi. Successful handover results with the better determination of handover points. So, an Artificial Neural Network-based network model to understand the network characteristics is firstly developed. Under simulated environment, the Received Signal Strength (RSS) of the heterogeneous network is observed to construct the training library. The trained network predicts RSS for resolving the handover points in the heterogeneous network. To ensure precise learning of the neural network about the RSS network characteristics, a renowned Whale Optimization Algorithm (WOA) is developed. The performance of WOA-NN model is compared with the conventional Levenberg–Marquardt-Neural Network, Fire Fly-Neural Network, Particle Swarm Optimization-Neural Network and Grey Wolf Optimization-Neural Network through throughput, handover, predicted RSS and Mean Absolute Error analyses. The predicted RSS of the proposed WOA-NN-based network model seems nearly closer to the actual model, attaining effective handoff.

Toward a Computational Historicism: From Literary Networks to the Autonomous Aesthetic

Stephen Greenblatt has identified pairs of moments in literary history such that the former moment must necessarily have preceded the later: literary history has a direction. This can be explained by asserting that the later texts required computational procedures capable of operating on the objects created by the earlier procedures, in the manner of Piaget’s reflective abstraction. Beyond Greenblatt’s examples two such pairs are examined, Amleth and Hamlet, The Winter’s Tale and Wuthering Heights. Also considered: Heuser and Le-Khac on 19th Century British novels. Three network-based network models are considered, at macro (topic modeling), meso (Moretti’s character networks), and micro scales (cognitive networks) of time.

N-NEMO: A Comprehensive Network Mobility Solution in Proxy Mobile IPv6 Network

In this paper, we propose a Network-based NEtwork MObility supporting scheme (N-NEMO) in Proxy Mobile IPv6 (PMIPv6) network, which is an issue still up in the air for the basic PMIPv6 protocol. The N-NEMO, like PMIPv6, bases mobility support on network functionality, thus enabling conventional (i.e., not mobility-enabled) IP devices to change their point of attachment without disrupting ongoing communications. As a result, N-NEMO enables off-the-shelf IP devices to roam within the fixed infrastructure, attach to a mobile network and move with it, and also roam between fixed and mobile points of attachment while using the same IP address. Besides, a tunnel splitting scheme is used in N-NEMO to differentiate the inter-Mobility Access Gateway (MAG) mobility and intra-MAG mobility. The analyzing results show that N-NEMO reduces the signaling cost significantly and enhances the efficiency and scalability of network mobility in the PMIPv6 context.