ResDNViT: A hybrid architecture for Netflow-based attack detection using a residual dense network and Vision Transformer DOI Creative Commons
Hassan Wasswa, Hussein A. Abbass, Timothy Lynar

и другие.

Expert Systems with Applications, Год журнала: 2025, Номер unknown, С. 127504 - 127504

Опубликована: Апрель 1, 2025

Язык: Английский

Deep-IDS: A Real-Time Intrusion Detector for IoT Nodes Using Deep Learning DOI Creative Commons
Sandeepkumar Racherla, Prathyusha Sripathi, Nuruzzaman Faruqui

и другие.

IEEE Access, Год журнала: 2024, Номер 12, С. 63584 - 63597

Опубликована: Янв. 1, 2024

The Internet of Things (IoT) represents a swiftly expanding sector that is pivotal in driving the innovation today's smart services. However, inherent resource-constrained nature IoT nodes poses significant challenges embedding advanced algorithms for cybersecurity, leading to an escalation cyberattacks against these nodes. Contemporary research Intrusion Detection Systems (IDS) predominantly focuses on enhancing IDS performance through sophisticated algorithms, often overlooking their practical applicability. This paper introduces Deep-IDS, innovative and practically deployable Deep Learning (DL)-based IDS. It employs Long-Short-Term-Memory (LSTM) network comprising 64 LSTM units trained CIC-IDS2017 dataset. Its streamlined architecture renders Deep-IDS ideal candidate edge-server deployment, acting as guardian between Denial Service (DoS), Distributed (DDoS), Brute Force (BRF), Man-in-the-Middle (MITM), Replay (RP) Attacks. A distinctive aspect this trade-off analysis intrusion detection rate false alarm rate, facilitating real-time Deep-IDS. system demonstrates exemplary 96.8% overall classification accuracy 97.67%. Furthermore, achieves precision, recall, F1-scores 97.67%, 98.17%, 97.91%, respectively. On average, requires 1.49 seconds identify mitigate attempts, effectively blocking malicious traffic sources. remarkable efficacy, swift response time, design, novel defense strategy not only secure but also interconnected sub-networks, thereby positioning IoT-enhanced computer networks.

Язык: Английский

Процитировано

7

MULTI-BLOCK: A novel ML-based intrusion detection framework for SDN-enabled IoT networks using new pyramidal structure DOI

Ahmed A. Toony,

Fayez Alqahtani, Yasser M. Alginahi

и другие.

Internet of Things, Год журнала: 2024, Номер 26, С. 101231 - 101231

Опубликована: Май 19, 2024

Язык: Английский

Процитировано

6

Securing the IoT Cyber Environment: Enhancing Intrusion Anomaly Detection With Vision Transformers DOI Creative Commons

Laraib Sana,

Mohsin Nazir, Jing Yang

и другие.

IEEE Access, Год журнала: 2024, Номер 12, С. 82443 - 82468

Опубликована: Янв. 1, 2024

ABSTRACT Background and Motivation: The ever-expanding Internet of Things (IoT) landscape presents a double-edged sword. While it fosters interconnectedness, the vast amount data generated by IoT devices creates larger attack surface for cybercriminals. Intrusions in these environments can have severe consequences. To combat this growing threat, robust intrusion detection systems (IDS) are crucial. comprised is multivariate, highly complex, non-stationary, nonlinear. extract complex patterns from data, we require most robust, optimized tools. Methods: Machine learning (ML) deep (DL) emerged as powerful tools IDSs, offering high accuracy detecting preventing security breaches. This research delves into anomaly detection, technique that identifies deviations normal system behavior, potentially indicating attacks. Given complexity explore methods to improve performance. Proposed Approach: investigates design evaluation novel IDS. We leverage optimize supervised ML like tree-based Support Vector Machines (SVM), ensemble methods, neural networks (NN) alongside cutting-edge DL approach long short-term memory (LSTM) vision transformers (ViT). hyperparameters algorithms using Bayesian optimization approach. Results: implemented models achieved impressive training accuracy, with Random Forest Ensemble Bagged Tree surpassing 99.90% an AUC 1.00, F1-score, balanced Matthews Correlation Coefficient (MCC) 99.78%. initial LSTM model yielded 99.97%, proposed ViT architecture significantly boosted performance 100% all metrics, along validation 78.70% perfect accuracy. Conclusion: Our findings demonstrate potential enhanced detection. improved capability be instrumental safeguarding integrity, identifying fraudulent activity, optimizing networks.

Язык: Английский

Процитировано

5

A Novel Data Preprocessing Model for Lightweight Sensory IoT Intrusion Detection DOI Creative Commons
Shahbaz Ahmad Khanday, Hoor Fatima, Nitin Rakesh

и другие.

International Journal of Mathematical Engineering and Management Sciences, Год журнала: 2024, Номер 9(1), С. 188 - 204

Опубликована: Янв. 14, 2024

IoT devices or sensor nodes are essential components of the machine learning (ML) application workflow because they gather abundant information for building models with sensors. Uncontrollable factors may impact this process and add inaccuracies to data, raising cost computational resources data preparation. Choosing best method pre-processing stage can lessen complexity ML wasteful bandwidth use cloud processing. Devices in ecosystem limited provide an easy target attackers, who make these create botnets spread malware. To repel attacks directed towards IoT, robust lightweight intrusion detection systems need hour. Furthermore, preprocessing remains first step modish models, ensemble techniques, hybrid methods developing anti-intrusion applications IoT. This article proposes a novel model as core structure using Extra Tree classifier feature selection two classifiers LSTM 1D-CNN classification. The dataset used research is CIC 2023 34 attack classes SMOTE (Synthetic Memory Oversampling Technique) has been class balancing. evaluates performance on 23 classification metrics. proposed approach obtained 92% accuracy 99.87% accuracy.

Язык: Английский

Процитировано

4

BGHOE2EB Model: Enhancing IoT Security With Gaussian Artificial Hummingbird Optimization and Blockchain Technology DOI Open Access

Kavitha Dhanushkodi,

Kiruthika Venkataramani,

N. R.

и другие.

Transactions on Emerging Telecommunications Technologies, Год журнала: 2025, Номер 36(1)

Опубликована: Янв. 1, 2025

ABSTRACT The Internet of Things (IoT) is transforming numerous sectors but also presents unique security challenges due to its interconnected and resource‐constrained devices. This study introduces the Bidirectional Gaussian Hummingbird Optimized End‐to‐End Blockchain (BGHO‐E2EB) model, designed detect classify cyberattacks within IoT environments. Unlike preventive approaches, developed model focuses on real‐time detection categorization attacks, enabling timely responses emerging threats. proposed integrates blockchain technology through Ethereum‐based smart contracts enhance integrity data exchanges networks. Additionally, a Artificial Algorithm employed for optimal feature selection, minimizing dimensionality computational load. A Long Short‐Term Memory (Bi‐LSTM) network further improves model's capability by accurately detecting categorizing cyber threats based selected features. Adam optimizer used efficient parameter tuning Bi‐LSTM network, ensuring high‐performance cyberattack detection. was evaluated using established benchmarks, including UNSW‐NB15, BOT‐IoT, NSL‐KDD datasets, accomplishing an accuracy 98.7%, precision 96.3%, level 99.5%, significantly outperforming traditional methods. These results demonstrate effectiveness BGHO‐E2EB as robust tool classifying in networks, making it suitable real‐world deployment dynamic environments where paramount.

Язык: Английский

Процитировано

0

An intrusion detection model based on Convolutional Kolmogorov-Arnold Networks DOI Creative Commons
Zhen Wang, Anazida Zainal, Maheyzah Md Siraj

и другие.

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Янв. 14, 2025

The application of artificial neural networks (ANNs) can be found in numerous fields, including image and speech recognition, natural language processing, autonomous vehicles. As well, intrusion detection, the subject this paper, relies heavily on it. Different detection models have been constructed using ANNs. While ANNs are relatively mature to construct models, some challenges remain. Among most notorious these bloated caused by large number parameters, non-interpretability models. Our paper presents Convolutional Kolmogorov-Arnold Networks (CKANs), which designed overcome difficulties provide an interpretable accurate model. (KANs) developed from representation theorem. Meanwhile, CKAN incorporates a convolutional computational mechanism based KAN. model proposed is incorporating attention mechanisms into CKAN's logic. datasets CICIoT2023 CICIoMT2024 were used for training validation. From results evaluating performance indicators experiments, CKANs has attractive prospect. compared with other methods, predict much higher level accuracy significantly fewer parameters. However, it not superior terms memory usage, execution speed energy consumption.

Язык: Английский

Процитировано

0

Secure and Efficient Lightweight Authentication Protocol (SELAP) for multi-sector IoT applications DOI
Amir Javadi, Sadegh Sadeghi, Peyman Pahlevani

и другие.

Internet of Things, Год журнала: 2025, Номер unknown, С. 101499 - 101499

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

0

Hybrid Neural Network-Based Intrusion Detection System: Leveraging LightGBM and MobileNetV2 for IoT Security DOI Open Access

Yeh-Fen Yang,

Ko-Chin Chang, Jia-Ning Luo

и другие.

Symmetry, Год журнала: 2025, Номер 17(3), С. 314 - 314

Опубликована: Фев. 20, 2025

The rapid expansion of the Internet Things (IoT) has uncovered a significant asymmetry in cybersecurity, where low-power edge devices must face sophisticated threats from adversaries backed by ample resources. In our study, we employ symmetry-based approach to rebalance these uneven scenarios. We propose Hybrid Neural Network Intrusion Detection System (Hybrid NNIDS) that uses LightGBM filter anomalies at traffic level and MobileNetV2 for further detection packet level, creating viable compromise between accuracy computational cost. Additionally, proposed NNIDS model, on ACI-IoT-2023 dataset, outperformed other intrusion models with an 94%, F1-score 91%, precision rate 93% attack detection. results indicate developed algorithm can greatly reduce processing overhead while still being able be implemented IoT environments. focus future work will real-world deployment security infrastructures their adaptation newer types vectors may malware.

Язык: Английский

Процитировано

0

Advanced protection technologies for microgrids: Evolution, challenges, and future trends DOI Creative Commons
Priya Ranjan Satpathy, Vigna K. Ramachandaramurthy, Sanjeevikumar Padmanaban

и другие.

Energy Strategy Reviews, Год журнала: 2025, Номер 58, С. 101670 - 101670

Опубликована: Март 1, 2025

Язык: Английский

Процитировано

0

A Novel Approach to IoT Device Identification via Anti‐Interference Dynamic Integral Neural Network and Multiobjective Fitness‐Dependent Optimizer Algorithm DOI Open Access

E. Anbalagan,

M. Kanchana,

G. Manikandan

и другие.

International Journal of Communication Systems, Год журнала: 2025, Номер 38(7)

Опубликована: Март 11, 2025

ABSTRACT The Internet of Things (IoT) has observed an accelerated development in the quantity applications due to rapid information technology. It can be difficult identify IoT devices heterogeneous, interference‐prone networks. accuracy, optimization, and robustness existing techniques are insufficient for dependable classification application detection. To overcome this complication, a novel approach device identification using anti‐interference dynamic integral neural network (AIDINN) multiobjective fitness‐dependent optimizer algorithm (MOFDOA) (IoT‐DTI‐AIDINN‐MOFDOA) is proposed. input data collected from Network Traffic Dataset. Then, given feature extraction. By synchro‐transient‐extracting transform (STET), features extracted dataset. Then AIDINN identification, which classifies known unknown devices. In general, does not adopt any optimization determine ideal parameters ensuring accurate identification. Hence, MOFDOA proposed here optimize AIDINN, precisely constructs performance measures like precision, recall, specificity, F measure, computational time, complexity evaluated. IoT‐DTI‐AIDINN‐MOFDOA method attains higher accuracy 25.23%, 16.12%, 21.27% precision 25.26%, 16.22%, 26.27% when analyzed with following models: type detection deep (IoT‐DTI‐DNN), adversarial attacks long short‐term memory (AA‐IoT‐LSTM), depending on fully connected (IoT‐DI‐FCNN), respectively.

Язык: Английский

Процитировано

0