An efficient convolutional neural network based attack detection for smart grid in 5G-IOT

International Journal of Critical Infrastructure Protection, Journal Year: 2025, Volume and Issue: unknown, P. 100738 - 100738

Published: Jan. 1, 2025

Language: Английский

---

Graph attention and Kolmogorov–Arnold network based smart grids intrusion detection

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: March 13, 2025

The digital revolution in power systems has increased their complexity and interconnectivity, thereby exacerbating the risk of cyberattacks. To protect critical infrastructure, there is an urgent need for advanced intrusion detection system capable capturing intricate interactions within smart grids. Although traditional graph neural network (GNN)-based methods have exhibited substantial potential, they primarily rely on data (e.g., IP addresses ports) to construct structure, failing adequately integrate physical from grid devices. Moreover, these typically employ fixed activation functions downstream deep networks, which limits accurate representation complex nonlinear attack patterns, reducing accuracy. address challenges, this paper introduces GraphKAN, a novel framework that leverages attention (GAT) Kolmogorov–Arnold (KAN) enhance precision GraphKAN firstly constructs graph-structured with devices, information technology communication devices as nodes, integrates connections logical dependencies among infrastructure elements edges, providing comprehensive view device interactions. Furthermore, GAT module utilizes multi-head mechanisms dynamically allocate node weights, extracting global features encompass both feature interaction patterns. KAN learnable based parameterized B-splines, enhancing expression extracted by significantly improving accuracy Experiments conducted datasets obtained Mississippi State University Oak Ridge National Laboratory demonstrate achieves accuracies 97.63%, 98.66%, 99.04% binary, ternary, 37-class tasks, respectively. These results represent improvements over state-of-the-art models, including GA-RBF-SVM, BGWO-EC, Net_Stack, gains 5.73%, 0.89%, 3.52%, findings underscore efficacy grids its robust performance scenarios.

Language: Английский

---

Adaptive protocols for hypervisor security in cloud infrastructure using federated learning-based anomaly detection

Engineering Applications of Artificial Intelligence, Journal Year: 2025, Volume and Issue: 152, P. 110750 - 110750

Published: April 15, 2025

Language: Английский

---

Attack detection and network recovery using Correlation Aware Lotus Effect Hierarchical Dual Graph Neural Networks

Computers & Electrical Engineering, Journal Year: 2025, Volume and Issue: 124, P. 110315 - 110315

Published: April 25, 2025

Language: Английский

---

A novel Bayesian optimizable ensemble bagged trees model for cryptocurrency fraud prediction approach

Security and Privacy, Journal Year: 2024, Volume and Issue: 7(6)

Published: May 19, 2024

Abstract Nowadays, the prediction of cryptocurrency side effects on critical aspects exchange rates in intelligent business is one main challenges financial market. Cryptocurrency defined as a set digital information concerning internal protocols marketing, such blockchain, which operates according to decentralized architecture. On other hand, fraud activities Ethereum transfer and management now increase affect safe transactional processes. This article presents new machine‐learning approach Detection based Bayesian Optimizable Ensemble Bagged Trees (BOEBT) algorithm. Moreover, goal this study derive accuracy model using different algorithms compare their evaluation parameters together. The performance proposed machine learning was evaluated by MATLAB tool. experimental results show that BOEBT algorithm merits achieving 99.21% 99.14% F1‐Score for prediction.

Language: Английский

---

DeepSecure Net: An Ensemble DPATMFNet approach with Enhanced Feature Selection for Advanced SDN Network Intrusion Detection System

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: Aug. 26, 2024

By enabling the control and administration of entire network from a single location, Software-Defined Network (SDN) was created to streamline administration. SDN controllers find intruders appealing because they make good targets. Attackers can take an controller use it route traffic according their requirements, which have disastrous effects on network. Although integrating with deep learning strategies opens up novel avenues for IDS deployment defense, detection models' efficacy depends quality training data. While non-identifiable systems (NIDSs) has yielded promising results recently several problems, most studies overlooked impact imbalanced redundant datasets. Therefore, improve intrusions via binary multiclass categorization, we proposed enhanced ensemble DL-based Dual Parallel Attention Transformer (DPAT) Modular Deep Fully Convolutional (MDFCN), termed DPATMFNet approach. An Enhanced AlexNet method extracts features input The Boosted Binary Meerkat Optimization Algorithm (BBMOA) is applied choose key features. system categorizes attacks, separates malicious benign traffic, identifies outstanding performance sub-attack types. Three current realistic datasets were used evaluation demonstrate effectiveness suggested system. We examined contrasted its that other IDSs. experimental findings indicate performs better than others at identifying various attacks. achieve accuracy, rate, precision above 99% compared existing approaches. show how effective model obtaining high accuracy while requiring shorter period.

Language: Английский

---

A Survey on 5G Wireless Network Intrusion Detection Systems Using Machine Learning Techniques

Advances in digital crime, forensics, and cyber terrorism book series, Journal Year: 2024, Volume and Issue: unknown, P. 195 - 218

Published: Dec. 30, 2024

This survey explores the landscape of Intrusion Detection Systems (IDS) in 5G wireless networks, with a specific emphasis on those leveraging Machine Learning (ML) techniques. The deployment is poised to revolutionize telecommunications unprecedented data rates, reduced latency, and enhanced capacity. However, advanced infrastructure also brings heightened security challenges, necessitating robust adaptive measures. IDS have emerged as key strategy counteract these vulnerabilities, particularly that exploit machine learning for anomaly detection rapid response. paper reviews several implementations across different domains, including Core Network, Internet Things (IoT), Smart Grids. Each domain's technical cybersecurity environment explored better context understanding available, which are scrutinised based ML models used, datasets training testing, architecture.

Language: Английский

---

SARM: A network State-Aware Adaptive Routing Mutation method for power IoT

Computer Networks, Journal Year: 2024, Volume and Issue: 255, P. 110889 - 110889

Published: Nov. 9, 2024

Language: Английский

---