Mitigating Energy Consumption in Heterogeneous Mobile Networks Through Data-Driven Optimization

IEEE Transactions on Network and Service Management, Journal Year: 2024, Volume and Issue: 21(4), P. 4369 - 4382

Published: June 19, 2024

Language: Английский

---

Distributed Denial of Services (DDoS) attack detection in SDN using Optimizer-equipped CNN-MLP

PLoS ONE, Journal Year: 2025, Volume and Issue: 20(1), P. e0312425 - e0312425

Published: Jan. 27, 2025

Software-Defined Networks (SDN) provides more control and network operation over a infrastructure as an emerging revolutionary paradigm in networking. Operating the many applications preserving services functions, SDN controller is regarded operating system of SDN-based architecture. The has several security problems because its intricate design, even with all amazing features. Denial-of-service (DoS) attacks continuously impact users Internet service providers (ISPs). Because centralized distributed denial (DDoS) on are frequent may have widespread effect network, particularly at layer. We propose to implement both MLP (Multilayer Perceptron) CNN (Convolutional Neural Networks) based conventional methods detect Denial Services attack. These models got complex optimizer installed them decrease false positive or DDoS case detection efficiency. use SHAP feature selection technique improve procedure. By assisting identification which features most essential spot incidents, approach aids process enhancing precision flammability. Fine-tuning hyperparameters help Bayesian optimization obtain best model performance another important thing that we do our model. Two datasets, InSDN CICDDoS-2019, utilized assess effectiveness proposed method, 99.95% for true (TP) CICDDoS-2019 dataset 99.98% dataset, results show highly accurate.

Language: Английский

---

Federated Learning Framework for Real-Time Activity and Context Monitoring Using Edge Devices

Sensors, Journal Year: 2025, Volume and Issue: 25(4), P. 1266 - 1266

Published: Feb. 19, 2025

With the increasing need for effective elderly care solutions, this paper presents a novel federated learning-based system that uses smartphones as edge devices to monitor and enhance in real-time. In system, individuals carry equipped with Inertial Measurement Unit (IMU) sensors, including an accelerometer activity recognition, barometer altitude detection, combination of accelerometer, gyrometer, magnetometer location tracking. The continuously collect real-time data go about their daily routines. These are processed locally on each device train personalized models recognition contextual monitoring. trained then sent server, where FedAvg algorithm is used aggregate model parameters, creating improved global model. This aggregated subsequently distributed back smartphones, enhancing capabilities. addition updates, information users' location, altitude, context server enable continuous monitoring tracking elderly. By integrating data, provides comprehensive framework supporting well-being across diverse environments. approach offers scalable efficient solution care, contributing enhanced safety overall quality life.

Language: Английский

---

LLM-AE-MP: Web attack detection using a large language model with autoencoder and multilayer perceptron

Expert Systems with Applications, Journal Year: 2025, Volume and Issue: unknown, P. 126982 - 126982

Published: Feb. 1, 2025

Language: Английский

---

A hybrid approach using support vector machine rule-based system: detecting cyber threats in internet of things

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Nov. 7, 2024

While the proliferation of Internet Things (IoT) has revolutionized several industries, it also created severe data security concerns. The these network devices and dependability IoT networks depend on efficient threat detection. Device heterogeneity, computing resource constraints, ever-changing nature cyber threats are a few obstacles that make detecting in systems difficult. Complex often go undetected by conventional measures, requiring more sophisticated, adaptive detection methods. Therefore, this study presents Hybrid approach based Support Vector Machines Rule-Based Detection (HSVMR-D) method for an all-encompassing to identifying IoT. HSVMR-D employs SVM categorize known unknown using attributes acquired from data. Identifying attack signatures patterns rule-based approaches improves efficiency without retraining adapting pre-trained models new contexts. Moreover, protecting vital infrastructure sensitive data, provides thorough adaptable solution improve posture deployments. Comprehensive experiment analysis simulation results compared baseline have confirmed proposed HSVMR-D. Furthermore, increased resilience completely novel changing threats, fewer false positives, improved accuracy all outcomes show work outperforms others. is helpful where primary objective secure environment when resources limited.

Language: Английский

---

Federated learning based fire detection method using local MobileNet

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Dec. 5, 2024

Fire is a dangerous disaster that causes human, ecological, and financial ramifications. Forest fires have increased significantly in recent years due to natural artificial climatic factors. Therefore, accurate early prediction of essential. While significant advancements been made traditional Deep Learning (DL) methods for fire detection, challenges remain accurately pinpointing recognizing regions, especially diverse large environments, prevent damage effectively. To address these challenges, this paper introduces novel Federated (FL)-based method called Indoor-Outdoor FireNet (IOFireNet) detecting localizing regions. The proposed incorporates Bilateral Filter (BF) effectively preprocess images reduce noise artifacts enhance detection clarity. It employs Super Pixel-based Adaptive Clustering (SPAC) precisely segment non-fire A global IOFireNet model developed aggregate parameters from local models, improving accuracy across varied while MobileNet used efficient data processing, enabling predictions on spread, severity, affected areas support warnings. FL-based attains an rate 98.65% 97.14% mean IoU segmentation. SPAC reaches 4.06%, which 2.45% better than the graph cut algorithm CRF model. achieves 0.23%, 4.20%, 3.29%, 10.02%, VGG-19, ResNet-50, Inception, Dense Net, respectively.

Language: Английский

---

Adaptive federated learning for resource-constrained IoT devices through edge intelligence and multi-edge clustering

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Nov. 20, 2024

In the rapidly growing Internet of Things (IoT) landscape, federated learning (FL) plays a crucial role in enhancing performance heterogeneous edge computing environments due to its scalability, robustness, and low energy consumption. However, one major challenges such is efficient selection nodes optimization resource allocation, especially dynamic resource-constrained settings. To address this, we propose novel architecture called Multi-Edge Clustered Edge AI Heterogeneous Federated Learning (MEC-AI HetFL), which leverages multi-edge clustering AI-driven node communication. This enables collaborate, dynamically selecting significant optimizing global tasks with complexity. Compared existing solutions like EdgeFed, FedSA, FedMP, H-DDPG, MEC-AI HetFL improves quality score, accuracy, offering up 5 times better distributed environments. The solution validated through simulations network traffic tests, demonstrating ability key IoT deployments.

Language: Английский

---

Neural network-based symmetric encryption algorithm with encrypted traffic protocol identification

PeerJ Computer Science, Journal Year: 2025, Volume and Issue: 11, P. e2750 - e2750

Published: April 4, 2025

Cryptography is a cornerstone of power grid security, with the symmetry and asymmetry cryptographic algorithms directly influencing resilience systems against cyberattacks. Cryptographic algorithm identification, critical component cryptanalysis, pivotal to assessing security hinges on core characteristics symmetric asymmetric encryption methods. A key challenge lies in discerning subtle spatial distribution patterns within ciphertext data infer underlying algorithms, which essential for ensuring communication systems. In this study, we first introduce plaintext guessing model (SCGM model) based leveraging strengths convolutional neural networks evaluate capabilities four algorithms. This assessed its learning efficacy practical applicability. We investigate protocol identification encrypted traffic data, proposing novel scheme that integrates temporal features. Special emphasis placed performance both frameworks. Experimental results demonstrate effectiveness our proposed scheme, highlighting potential enhancing security.

Language: Английский

---

Assembling multipath service function chains in substrate graphs using sharing instances and deep learning

Engineering Applications of Artificial Intelligence, Journal Year: 2025, Volume and Issue: 152, P. 110900 - 110900

Published: April 16, 2025

Language: Английский

---

Allocation of computing resources based on multi-objective strategy and performance improvement in 5G networks

Computer Communications, Journal Year: 2025, Volume and Issue: unknown, P. 108197 - 108197

Published: April 1, 2025

Language: Английский

---