Hybrid deep representation learning-based adversarial attack detection framework for secure communication in 6G-assisted IoT systems

Mohammed A. AlAqil; Amal K. Alkhalifa; Sultan Alahmari; Nadhem NEMRI; Saied Alshahrani; Sulaiman Alamro; Sultan Almutairi; Achraf Ben Miled; Mohammed A. AlAqil; Amal K. Alkhalifa; Sultan Alahmari; Nadhem NEMRI; Saied Alshahrani; Sulaiman Alamro; Sultan Almutairi; Achraf Ben Miled

doi:10.3934/math.20261258

AIMS Mathematics

2026, Volume 11, Issue 9: 31990-32019. doi: 10.3934/math.20261258

Previous Article Next Article

Research article Special Issues

Hybrid deep representation learning-based adversarial attack detection framework for secure communication in 6G-assisted IoT systems

1.
Department of Electrical Engineering, College of Engineering, King Faisal University, AlAhsa, Saudi Arabia
2.
Department of Computer Sciences, College of Computer and Information Sciences, Princess Nourah bint Abdulrahman University, P.O. Box 84428, Riyadh 11671, Saudi Arabia
3.
King Abdul Aziz City for Science and Technology (KACST), Cybersecurity Institute, Riyadh, Saudi Arabia
4.
Department of Information Systems, Applied College at Mahayil, King Khalid University, Saudi Arabia
5.
Department of Computer Science and Artificial Intelligence, College of Computing and Information Technology, University of Bisha, Bisha 61922, Saudi Arabia
6.
Department of Computer Science, College of Computer, Qassim University, Buraydah 51452, Saudi Arabia
7.
Department of Computer Science, Applied College, Shaqra University, Shaqra 15526, Saudi Arabia
8.
Department of Computer Science, College of Science, Northern Border University, Arar 73213, Saudi Arabia

Received: 05 April 2026 Revised: 29 July 2026 Accepted: 31 July 2026 Published: 28 September 2026
MSC : 68T07, 68T45

Internet of Things (IoT) networks play a significant role in sixth-generation (6G) networks due to higher throughput, lower latency, ultra-reliability, and dense connectivity. Artificial intelligence (AI), particularly deep learning (DL) and machine learning (ML), offers an effective means to advance and implement next-generation radio communication technologies. However, these approaches are vulnerable to adversarial attacks, which can significantly degrade system performance and lead to inaccurate predictions. This outcome is highly undesirable for a reliable and ubiquitous IoT-based communication system. This paper presents a hybrid neural network framework for adversarial machine learning attack detection using ridge regression (HNNF-AMLADRR) in 6G network-assisted intelligent IoT systems. The main objective is to develop an effective detection framework for adversarial machine learning (AML) attacks, thereby improving the security of intelligent systems in 6G and IoT environments against malicious perturbations. First, the min-max scaler normalization method is employed in data preprocessing to transform the input data into a suitable format. Moreover, ridge regression is primarily used for feature reduction to remove irrelevant or redundant features. For adversarial attack detection, the proposed HNNF-AMLADRR model designs a hybrid model of a self-attention-based restricted Boltzmann machine with a recurrent convolutional neural network technique. Finally, the model parameters are fine-tuned using the Adafactor optimizer to enhance classification performance. The experimental evaluation of the HNNF-AMLADRR algorithm was done using the AML dataset. Empirical outcomes indicated an enhanced performance of the HNNF-AMLADRR model compared to recent approaches.
- adversarial attacks,
- machine learning,
- ridge regression,
- sixth-generation network,
- Internet of Things,
- self-attention
Citation: Mohammed A. AlAqil, Amal K. Alkhalifa, Sultan Alahmari, Nadhem NEMRI, Saied Alshahrani, Sulaiman Alamro, Sultan Almutairi, Achraf Ben Miled. Hybrid deep representation learning-based adversarial attack detection framework for secure communication in 6G-assisted IoT systems[J]. AIMS Mathematics, 2026, 11(9): 31990-32019. doi: 10.3934/math.20261258

Related Papers:

Abstract

Internet of Things (IoT) networks play a significant role in sixth-generation (6G) networks due to higher throughput, lower latency, ultra-reliability, and dense connectivity. Artificial intelligence (AI), particularly deep learning (DL) and machine learning (ML), offers an effective means to advance and implement next-generation radio communication technologies. However, these approaches are vulnerable to adversarial attacks, which can significantly degrade system performance and lead to inaccurate predictions. This outcome is highly undesirable for a reliable and ubiquitous IoT-based communication system. This paper presents a hybrid neural network framework for adversarial machine learning attack detection using ridge regression (HNNF-AMLADRR) in 6G network-assisted intelligent IoT systems. The main objective is to develop an effective detection framework for adversarial machine learning (AML) attacks, thereby improving the security of intelligent systems in 6G and IoT environments against malicious perturbations. First, the min-max scaler normalization method is employed in data preprocessing to transform the input data into a suitable format. Moreover, ridge regression is primarily used for feature reduction to remove irrelevant or redundant features. For adversarial attack detection, the proposed HNNF-AMLADRR model designs a hybrid model of a self-attention-based restricted Boltzmann machine with a recurrent convolutional neural network technique. Finally, the model parameters are fine-tuned using the Adafactor optimizer to enhance classification performance. The experimental evaluation of the HNNF-AMLADRR algorithm was done using the AML dataset. Empirical outcomes indicated an enhanced performance of the HNNF-AMLADRR model compared to recent approaches.

References

[1]	R. Kumar, J. Dutta, N. Vamsi, U. S. Varri, D. Puthal, Next-generation security in the 6G era: The role of AI in safeguarding future networks, IEEE Access, 14 (2026), 17347–17380. https://doi.org/10.1109/ACCESS.2025.3650208 doi: 10.1109/ACCESS.2025.3650208
[2]	A. Tripathi, A. S. Anagha, A. S. Kumar, S. Anjankar, S. Balpande, S. W. Prakash, Enhancing cybersecurity resilience in 6G networks using adversarial machine learning, In: Emerging perspectives and applications of computational intelligence and smart systems, CRC Press, 2025,338–344.
[3]	R. Huang, Y. C. Li, Adversarial attack mitigation strategy for machine learning-based network attack detection model in power system, IEEE Trans. Smart Grid, 14 (2023), 2367–2376. https://doi.org/10.1109/TSG.2022.3217060 doi: 10.1109/TSG.2022.3217060
[4]	S. Chen, M. H. Xue, L. L. Fan, S. Hao, L. H. Xu, H. J. Zhu, et al., Automated poisoning attacks and defenses in malware detection systems: An adversarial machine learning approach, Comput. Secur. , 73 (2018), 326–344. https://doi.org/10.1016/j.cose.2017.11.007 doi: 10.1016/j.cose.2017.11.007
[5]	V. R. Kebande, S. Alawadi, F. M. Awaysheh, J. A. Persson, Active machine learning adversarial attack detection in the user feedback process, IEEE Access, 9 (2021), 36908–36923. https://doi.org/10.1109/ACCESS.2021.3063002 doi: 10.1109/ACCESS.2021.3063002
[6]	Y. L. Wang, T. Sun, S. H. Li, X. Yuan, W. Ni, E. Hossain, et al., Adversarial attacks and defenses in machine learning-empowered communication systems and networks: A contemporary survey, IEEE Commun. Surv. Tutor. , 25 (2023), 2245–2298. https://doi.org/10.1109/COMST.2023.3319492 doi: 10.1109/COMST.2023.3319492
[7]	F. Mumcu, K. Doshi, Y. Yilmaz, Adversarial machine learning attacks against video anomaly detection systems, In: 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2022,205–212. https://doi.org/10.1109/CVPRW56347.2022.00034
[8]	M. Standen, J. Kim, C. Szabo, Adversarial machine learning attacks and defences in multi-agent reinforcement learning, ACM Comput. Surv. , 57 (2025), 1–35. https://doi.org/10.1145/3708320 doi: 10.1145/3708320
[9]	A. Sharma, S. Rani, Enhancing 6G-IoT network security: A trustworthy and responsible AI-driven stacked-hybrid model for attack detection, IEEE Int. Things J. , 13 (2026), 7777–7784. https://doi.org/10.1109/JIOT.2025.3566403 doi: 10.1109/JIOT.2025.3566403
[10]	H. Y. Lin, B. Biggio, Adversarial machine learning: attacks from laboratories to the real world, Computer, 54 (2021), 56–60. https://doi.org/10.1109/MC.2021.3057686 doi: 10.1109/MC.2021.3057686
[11]	F. O. Catak, S. Y. Yayilgan, Deep neural network based malicious network activity detection under adversarial machine learning attacks, In: International Conference on Intelligent Technologies and Applications, Cham: Springer, 2021,280–291. https://doi.org/10.1007/978-3-030-71711-7_23
[12]	A. V. Ribeiro, A. L. R. Madureira, L. N. Sampaio, Secure by design: Merging network and security bootstrapping for IoT systems through NDN, Comput. Netw. , 280 (2026), 112162. https://doi.org/10.1016/j.comnet.2026.112162 doi: 10.1016/j.comnet.2026.112162
[13]	H. Shen, J. C. Wu, J. Wang, Z. G. Wu, Adversarial dynamic games for Markov jump systems: A policy iteration Q-learning method, Automatica, 183 (2026), 112591. https://doi.org/10.1016/j.automatica.2025.112591 doi: 10.1016/j.automatica.2025.112591
[14]	H. Shen, C. J. Peng, H. C. Yan, S. Y. Xu, Data-driven near optimization for fast sampling singularly perturbed systems, IEEE Trans. Automat. Control, 69 (2024), 4689–4694. https://doi.org/10.1109/TAC.2024.3352703 doi: 10.1109/TAC.2024.3352703
[15]	H. Ke, J. Xu, Y. Wang, H. Y. Chen, Z. P. Shen, Adversarial machine learning in cybersecurity: Attacks and defenses, Int. J. Manag. Sci. Res. , 8 (2025), 26–33. https://doi.org/10.53469/ijomsr.2025.08(02).04 doi: 10.53469/ijomsr.2025.08(02).04
[16]	S. R. Bommana, S. Veeramachaneni, S. E. Ahmed, M. B. Srinivas, Addressing adversarial attacks in IoT using deep learning AI models, IEEE Access, 13 (2025), 50437–50449. https://doi.org/10.1109/ACCESS.2025.3552529 doi: 10.1109/ACCESS.2025.3552529
[17]	K. Barik, S. Misra, L. Fernandez-Sanz, Adversarial attack detection framework based on optimized weighted conditional stepwise adversarial network, Int. J. Inform. Secur. , 23 (2024), 2353–2376. https://doi.org/10.1007/s10207-024-00844-w doi: 10.1007/s10207-024-00844-w
[18]	X. W. Yuan, S. Han, W. Huang, H. L. Ye, X. L. Kong, F. Zhang, A simple framework to enhance the adversarial robustness of deep learning-based intrusion detection system, Comput. Secur. , 137 (2024), 103644. https://doi.org/10.1016/j.cose.2023.103644 doi: 10.1016/j.cose.2023.103644
[19]	G. Petihakis, A. Farao, P. Bountakas, A. Sabazioti, J. Polley, C. Xenakis, AIAS: AI-assisted cybersecurity platform to defend against adversarial AI attacks, In: ARES '24: Proceedings of the 19th International Conference on Availability, Reliability Security, 2024, 1–7. https://doi.org/10.1145/3664476.3669920
[20]	Z. A. Sheikh, Y. Singh, P. K. Singh, P. J. S. Gonçalves, Defending the defender: Adversarial learning based defending strategy for learning based security methods in cyber-physical systems (CPS), Sensors, 23 (2023), 5459. https://doi.org/10.3390/s23125459 doi: 10.3390/s23125459
[21]	B. A. Alabsi, M. Anbar, S. D. A. Rihan, Conditional tabular generative adversarial based intrusion detection system for detecting DDoS and DoS attacks on the Internet of Things networks, Sensors, 23 (2023), 5644. https://doi.org/10.3390/s23125644 doi: 10.3390/s23125644
[22]	L. Dhamija, U. Bansal, AFLF: A defensive framework to defeat multi-faceted adversarial attacks via attention feature fusion, Evolv. Syst. , 16 (2025), 20. https://doi.org/10.1007/s12530-024-09643-z doi: 10.1007/s12530-024-09643-z
[23]	B. R. Cherukuri, Advanced multi class cyber security attack classification in IoT based wireless sensor networks using context aware depthwise separable convolutional neural network, J. Mach. Comput. , 5 (2025), 814–830. https://doi.org/10.53759/7669/jmc202505064 doi: 10.53759/7669/jmc202505064
[24]	C. S. Kodete, K. B. Raju, K. Karmakonda, S. Sikindar, J. V. N. Ramesh, N. S. K. M. K. Tirumanadham, Optimizing intrusion detection with triple boost ensemble for enhanced detection of rare and evolving network attacks, Int. J. Electr. Electron. Eng. Telecommun. , 14 (2025), 115–129. https://doi.org/10.18178/ijeetc.14.3.115-129 doi: 10.18178/ijeetc.14.3.115-129
[25]	Y. F. Zhang, X. Z. Gao, X. Peng, J. Q. Ye, X. Li, Attention-based recurrent temporal restricted Boltzmann machine for radar high resolution range profile sequence recognition, Sensors, 18 (2018), 1585. https://doi.org/10.3390/s18051585 doi: 10.3390/s18051585
[26]	D. Altınel, Development of deep learning optimizers: Approaches, concepts, and update rules, 2025, arXiv: 2509.18396.
[27]	CNR-IEIIT, Adversarial Machine Learning Dataset, Kaggle, 2022. Available from: https://www.kaggle.com/datasets/cnrieiit/adversarial-machine-learning-dataset.
[28]	I. Vaccari, A. Carlevaro, S. Narteni, E. Cambiaso, M. Mongelli, eXplainable and reliable against adversarial machine learning in data analytics, IEEE Access, 10 (2022), 83949–83970. https://doi.org/10.1109/ACCESS.2022.3197299
[29]	A. S. Albahri, R. A. Hamid, A. R. Abdulnabi, O. S. Albahri, A. H. Alamoodi, M. Deveci, et al., Fuzzy decision-making framework for explainable golden multi-machine learning models for real-time adversarial attack detection in vehicular ad-hoc networks, Inform. Fusion, 105 (2024), 102208. https://doi.org/10.1016/j.inffus.2023.102208 doi: 10.1016/j.inffus.2023.102208
[30]	C. Joshi, J. Kumar, G. Kumawat, Detection of unseen malware threats using generative adversarial networks and deep learning models, Sci. Rep. , 15 (2025), 34804. https://doi.org/10.1038/s41598-025-18811-3 doi: 10.1038/s41598-025-18811-3

Reader Comments

Your name:*

Email:*
© 2026 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)