Security Technologies for 6G Mobile Systems and Challenges Associated with AI Technologies

Arvinder Singh (Assistant Professor)

University Institute of Computing (UIC)
Chandigarh University, Mohali, Punjab
[email protected]
Jagvinder singh
University Institute of Computing (UIC)
Chandigarh University, Mohali, Punjab
[email protected]

ABSTRACT: The forthcoming network generation, 6G is intimately linked to the structuring and administration of intelligent
network, whereas 5G is known for its microservices-based architecture and cloud features. Consequently, Artificial intelligence (AI)
plays a pivotal role in the 6G framework. Nonetheless, the integration of 6G with AI can be a double-edged sword, as it has the
potential to either enhance or compromise security and privacy. The complete automation of future networks necessitates the
proactive identification of threats, the execution of intelligent mitigation strategies, and the assurance of self-reliant network in 6G.
Hence the utilization of AI (Artificial intelligence) in 6G security along with the potential issues and solutions to bolster the
contribution of AI in securing 6G network.

Index: Artificial Intelligence (AI), Machine Learning (ML), Intelligent Security (IS), 6G and Security of 6G.

I. INTRODUCTION
The disparity in how people connect with digital virtual worlds after 2030 can be predicted by the upcoming 6G communication era.
To generate digital virtual words with connected intelligence, future networks need to be equipped with the latest technologies.
Research indicates that although traditional uses such as multimedia streaming will continue, emerging application areas for 6G
systems include multisensory augmented reality (XR) applications, wireless brain-computer interaction. Other applications of 6G
include in-body networks for eHealth and holographic telepresence, which demand ultra-high data throughput, ultra-low latency, and
ultra-reliability [2]. A variety of dense networks combined with hyper-connection and intelligent connectivity are further features of
6G network. Th complex requirements of 6G networks require advanced security measures. Security built for 5G need to be further
enhanced using SDN and NFV ideas to meet the needs of 6G in terms of security [9].

After the release of the 5G system, research on the upcoming 6G network had begun. The Ten-year cycle norm in the mobile industry
is shown in the figure 1, which shows when the 6G rollout is expected to begin in 2030. By 2030 it is estimated that the monthly
traffic capacity of each user will be 257GB. This traffic has increased 50 times compared to 2020[10],[11]. Additionally, compared to
2020, the number of machine-to-machine (M2M) subscriptions will increase 422 times by 2030 [10],[12]. Nevertheless, there are
many difficulties and problems in dealing with this band (THZ), such as high path loss to FSPL and atmospheric absorption,
restriction on the distance of transmission links, hardware limitations, etc.
The entire 6G objective can be summarized in four keywords: “Deep connectivity”, “Intelligent connectivity”, “ubiquitous
connectivity” and “Holographic connectivity” [4],[6] and [7]. Free space optical (FSO) backhaul, which requires microsecond (μs)
latency and terabits per second (Tbsp.) data speeds, is used for these types of applications in place of optical fibre [8]. Nevertheless,
the operating bandwidth of the system must be increased to meet these requirements. Therefore, a significant amount of bandwidth
resources will be made accessible to support such applications be establishing terahertz slots (0.1 THz to 10 THz) to future 6G
systems [2],[3]. Hear the security concerns related to 6G networks are discussed. The security flaws that exist in 5G and 6G are
explained by the security that existed before 6G. Security concern arising from architectural modifications and new technologies
employed in 6G are also covered in the section. Possible attacks on different 6G infrastructure layers are depicted in figure 2.

A. Security Prior to 6G

5G network software technologies such as network function virtualization, network slicing, multi-access edge computing and software
defined networking may be integrated into 6G, which means they may have security flaws and will be fixed in 6G they might still
exist. Attacks on SDN controllers, vulnerabilities in the platforms used to deploy SDN controllers and applications and north and
south interfaces are the primary security issues related to SDN [10]. Virtual machines (VMs), virtual network functions (VNFs)
hypervisors, virtual network function manages, and network function orchestrators have been identified as NFV-related security issues
[11]. Because 6G systems are highly decentralized, 6G is vulnerable to MEC attacks, distributed denial of services (DDoS), and
physical security threats [12]. Network hacking can expose DoS attacks and data theft through compromised parts [13].
Fig. 1. Development of Mobile System Technologies.

Because resource-contained Internet of Things (IoT) devices cannot implement modern encryption techniques for strong security, they
are quite vulnerable to attacks [15]. It exploited; these gadgets can be used as weapons in cyber-attacks. Data collected by IoE using
6G applications is unsafe due to data security issues. Identity security, location privacy, and data privacy can all be jeopardized by
data breaches involving resource-limited IoT devices. Local 5G network are commonly used to provide services in various industries,
including heath care and education. This is further enhanced by 6G technology, which make it possible for embedded network, drone
swarms, and environment sensor network-smaller network with longer battery life. When necessary, these local networks can
collaborate with wide area networks while functioning as independent networks. Different embedded security levels are implemented
by multiple stakeholders in local 6G networks, which differ from well-defined local 5G networks. Attackers have the chance to launch
attacks on a local 6G network with weak security. and then go into the networks that are linked to the network compromised.

B. 6G Framework Security Network

In 6G network, small cells have all the characteristics of a mesh network, including extended cell deployment, multiple
connectivity options device to device (D2D) communications. The increasing number of vulnerable devices on dispersed networks
increases the threat surface and increase the potential for malicious entities to attack system. Subset definitions must be taken into
account when designing security protocols. A wide area network cannot safely protect the large number of devices inside each subset.
In 6G network it would more efficient to use a hierarchical security solution that separates broadband security from subnet traffics.
Higher-layer RAN services, such as User-Level Microservices (UPMS) and Control-Level Microservices (CPMS), became more
centralized under the RAN-core architecture and can even be deployed at the periphery of core operations. Attackers may target these
microservice’s, which oversee numerous radio units, impacting the network as a whole. Zero touch network and service management
architecture frameworks are used in 6G network to minimize human error, enable quick service launch, and save operating costs.
Nevertheless, these network’ complete automation and capacity for self-learning also leave them open to intricately interwoven attack
chains. Data protections become more difficult to ensure as zero-touch networks with little to no human intervention become
essential. The full automation of 6G network necessitates thinking about the integration of automated machine ethics.

C. 6G Technology Security

Artificial intelligence is essential to 6G networks being fully autonomous. As a result, 6G is vulnerable to machine learning (ML)
and other AI system threats. tainting attacks, data injection, stewardship, logical fallacies, model theft, model inversion, model
ancestry and membership inference and examples of possible security vulnerabilities in machine learning systems [16]. As AI systems
gather more data, they will be able to function more efficiently. Since data processing is frequently opaque to users, problems with
data gathering attacks and unintentional personal data usage add to data protection concerns. Furthermore, blockchain is an important
technology that can improve 6G systems' performance. Within the broad and decentralized domain of 6G networks, it can be used for
distributed resource management, spectrum allocation, and service management [17].
Blockchain instability can be achieved through 51% attacks using quantum computing. Privacy preservation presents a challenge
since blockchain networks store data in a publicly accessible format. Since the existing 5G standard relies on conventional
cryptography like Elliptic Curve Cryptography (ECC), it does not address the security risks associated with quantum computing. As
quantum computers become more prevalent in the 6G future, current security measures that rely on asymmetric key cryptography are
susceptible to attacks by these machines [18].
Asymmetric key encryption might not be able to adequately safeguard 5G communications from quantum threats in the absence of a
quantum-secure encryption method. Applications including indoor positioning systems and vehicle-to-vehicle communication are
suitable for visible light communication (VLC). However, common attacks like jamming, eavesdropping, and node compromise
techniques pose a threat to VLC's security [19].

Fig. 2: Intelligent 6G Architecture [6], and 6G Security and Privacy Issues.

II 6G APPLICATIONS AREA

5G Network won’t be able to support new intelligent and automated system in ten years, nor will they be able to keep up
with the data center and automated system that are growing at an accelerated rate in the future years. However, 6G has a lot of
potential for future cutting-edge applications because AI is fully incorporated into 6G communication network.

Below is an illustration of the primary application areas for 6G wireless communication:

 Extended reality (XR): Communication systems powered by 6G will support AR, VR, and MR. VR content, often known
as digital reality, completely submerges the user in an artificial
environment and is produced through digital technologies. AR is the overlaying of virtual items onto the real world. MR blends VR
and AR. XR has uses in a variety of industries, including entertainment, healthcare, and education. The 6G systems must operate in
real time in order to provide a realistic XR experience [20].

 Autonomous systems and robots: The use of autonomous systems and robots in 6G wireless technology will transform daily life.
One excellent example is the employment of sensor such as GPS, light detection, sonar, radar and an odometer by self-driving car to
sensor their surroundings.

 Super intelligent society: 6G infrastructure can be used to build an intelligent society, which will improve people's quality of life.
Artificial Intelligence and energy harvesting will be used to automate and monitor the environment. For instance, any business that
owns self-driving automobiles and smart mobile devices would be able to operate everything remotely. Smart houses will become a
reality once we get to the point where we can operate any piece of equipment from a distance.

 Brain computer Interaction (BCI): It is based on the idea that signals can be directly transmitted from the brain to digital devices.
With the use of 6G technology, these signals will be extracted, digitally processed, and translated into actions in real-time [43].

 Smart healthcare: The health sector, particularly the such as remote surgery and healthcare services, will be greatly assisted by 6G,
which is built on AI, 3D imaging, and mobile computing. The foundation of the real time nourishing of a database system through a
wireless communication system will be the future health system. This will allow for the sharing of a massive amount of health
information, which will directly improve the quality of healthcare.
 Automation and manufacturing: The industrial internet of everything (IIoE) will emerge from the internet of Everything (IoE) and
play a major role in the industrial field thanks to 6G technology as the fifth industrial revolution draws closer. By use of embedded
smart sensors, IIOE will be able to obtain vast amounts of data pertaining to diverse elegant industrial equipment and lines of
production. The needs of the technical pathways will determine how this data is examined, handled, and directed. As we approach the
6G era, billions of gadgets and pieces of equipment will be linked to the network, making the concept of smart industries a reality
[22].

III AN AI FOR 6G SECURITY AND PRIVACY

Artificial intelligence contributes the security of the security system. 6G systems. This section covers the application of AI in 6G
privacy, pre-6G security, 6G architecture security, and 6G technology security. Fig. 3: AI safeguards 6G applications.

A. Using AI to Recognize and Address Pre-6G Security Concerns

Networks with SDN/NFV enabled can effectively use Deep Neural Networks (DNN) and deep reinforcement learning for
multilayered intrusion detection and prevention [23]. In comparison with several conventional ways, they efficiently fight against
attacks such as IP spoofing, flow table overloading, DDoS, control plane saturation, and host location hijacking. Because of their
accuracy and speed of processing, machine learning techniques like Random Forest and Decision Trees have also shown promise in
identifying DDoS assaults in SDN systems [24]. As 6G networks anticipate the dynamic insertion of virtual functions on-demand,
Machine learning Attacks against SDN/NFV can be successfully repelled with learning-based adaptive security approaches. Using AI
approaches to identify vulnerabilities in a widely dispersed network, the attacks also adapt. Thus, detection methods that rely on rules
are not effective.

The traditional authentication/authorization solutions are ineffective for providing acceptable security in large-scale IoT due to on-
device reference restrictions, difficulty managing keys in highly diverse network and the abundance of device data creation. Industrial
Internet of Things (IIoT) peculiarity-based intrusion detection systems [25]. Malicious packets can be recognized by tracking their
movements. Zero-day threat identification is made possible by these learning-based detection systems, which take into account a
range of data properties as input. For devices with limited resources in the future networks, a better method is to use machine learning
in conjunction with communication link properties [26]. In this manner, gadgets avoid making use of their constrained resources to
offer more sophisticated protection.

In 6G, which can be thought of as an expansion of local 5G networks outside of vertical domains, learning-based security solutions
can be beneficial both within and between subnetworks. When ML-based algorithms are implemented at the perimeter, they are able
to monitor the activities of neighbouring subnetworks and identify any malicious activity coming from them. Since subnetworks often
function independently, large-scale data transfers between them may be useless. For the sake of effective communication, a sub-
network may only impart newly acquired security intelligence to another. A second subnetwork can apply dynamic policies, identify
malicious traffic on other networks, and use the shared intelligence to input into its machine learning models.

B. Using AI to Reduce Security Concerns with 6G Technologies

AI-powered predictive analytics is able to foresee attacks like blockchain 51% attacks before they happen. There could be a risk to
asymmetric key cryptography from a quantum computer. Nonetheless, they might give AI/ML algorithms exponential speed boosts,
enabling them to complete jobs far more quickly and accomplish previously unachievable goals. Therefore, one potential defence
strategy against assaults based on quantum computers is the use of quantum machine learning in network security [27]. Intelligent
beamforming provides the best beamforming policy for VLC system to prevent eavesdropper attacks.
beamforming approaches based on Reinforcement Learning (RL) [28]. Since DoS attacks and jamming are similar, peculiarity-based
detection systems with AI capabilities may be able to identify jamming attempts. AI-based authentication and authorization systems
are also suitable for resisting node compromise threats.

C. 6G Privacy Using AI
In 6G, several base stations enable simultaneous device communication through multiple connectivity mesh networks with small
cells. Devices could be able to transport data via privacy-preserving channels founded on a ranking of the routes that are dynamically
detected by edge-based machine learning algorithms. Compared to cloud-based integrated learning, federated learning maintains data
closer to the user in order to improve location and data privacy. Sixth generation sub-network level AI allows privacy preservation
within the sub-network and shares just the learned intelligence externally, hence reducing privacy concerns. For applications such as
in-body networks, data confinement within the network is appropriate. Users will prefer varying privacy levels on different
applications due to the large variety of 6G applications and the substantial data collection needed to feed intelligent models. Fully
automated 6G networks with privacy preservation may be supported by AI-mechanism based serviceability’s privacy-preserving
policy updates [29].

IV ISSUES AND POSSIBLE ACTIONS

Even if AI is crucial to 6G systems, there are still ethical, security, and privacy concerns with it. Furthermore, AI mechanism can used
as a device to carry out cunning attacks. The ethical, security, and privacy concerns with AI are discussed in this section along
with possible solutions. Figure 3 further demonstrates AI's function as a 6G application protector.

Security Concerns with AI

Issues: In particular, with ML systems susceptible to security flaws, 6G achieves linked intelligence through AI-enabled functions.
Staining attacks have an impact on a machine learning system's phase of learning, leading to inaccurate model learning. Data
injection, manipulation, and logical corruption are a few instances of poisoning attacks. Evasion attacks use deliberately constructed
hostile case to try and get around the model during the inference phase. Model reversal, membership inference, and model extraction
are some of the attacks on ML models that are based on APIs [16].

Solutions: It is possible to build resilient AI systems through countermeasures like hostile machine learning and moving target
defence. Additional defence strategies include defensive distillation and adversarial training against evasion attacks, input
confirmation and robust learning against Staining attacks, homomorphic encryption and unequal privacy against API-based assaults,
and so on. Designing these defence mechanisms presents a difficulty in striking a balance between enhanced defence and performance
deterioration [16].

Privacy Issues in AI

Issues: Artificial Intelligence (AI) has the potential to damage privacy due to its large-scale data analysis capabilities, future
computing speeds, and automation requirements for networks. The way that external systems manage users’ data is no longer within
their control, and billions of devices must gather an enormous amount of user data in order to support 6G. For instance, the suggested
intelligent authentication methods may employ private user data and are dependent on physical characteristics [26]. Data theft could
occur from unreliable IoT devices (such as weak sensors) that provide personal information to AI systems. In addition, privacy
violations may arise from model inversion attacks on machine learning to recover training data [30].

Solutions: Several privacy strategies that add random noise to training data are among the solutions; these helps keep private
information from being revealed to training models [32]. Blended learning with an edge preserves data closeness to the user through
the use of physical controllers [30]. Homomorphic encryption offers a technological privacy protection by enabling mathematical
computations on encrypted data [31]. To confirm that learning using encrypted data produces results similar to learning using
unencrypted data, more research is required.

AI-Based Threats to 6G Network

Issues: Through the use of a distributed edge-based architecture, artificial intelligence (AI) is able to identify patterns in vast amounts
of data at several positions (Smart beam, edge, and cloud) and make intelligent judgments across the network. Thus, network
vulnerabilities can be identified by AI-based techniques. For instance, AI is capable of identifying the most susceptible Internet of
Things devices, transforming them into automatons, and launching DDoS attacks [16] upside a vital node.

Solutions: Deploying smart defence system series as counterweight to intelligent AI-based attacks. The distributed intelligence
inherent in AI can strengthen these systems. Moving target defence techniques bring mobility to the network, disrupting the learning
capabilities AI based equipped attackers, thus providing a proactive defence [33]. Additionally, quantum

machine learning can develop sophisticated defence strategies against AI-based attacks [27].

V. CONCLUSIONS

The upcoming 6G mobile network are largely made possible by artificial intelligence, and understanding security is critical to
realizing the 6G goal. Artificial intelligence enable 6G data security provides robust and intelligent data security solutions. This paper
provides an overview of the advantages and difficulties of integrating intelligent security and privacy with artificial intelligence in 6G
system. Furthermore, it highlights the obstacles in providing AI-based security and privacy offers workable practical solutions that
suggest future inquiries.

REFERENCES

[1] A. M. Ahmed, S. A. Hasan, and S. A. Majeed, “5G mobile systems, challenges and technologies: A survey,” J. Theor. Appl. Inf.
Technol., vol. 97, no. 11, pp. 3214-3226, 2019.
[2] I. F. Akyildiz, A. Kak, and S. Nie, “6G and beyond: The future of wireless communications systems,” IEEE Access, vol. 8, pp.
133995–134030, 2020.
[3] W. Saad, M. Bennis, and M. Chen, “A vision of 6G wireless systems: Applications, trends, technologies, and open research
problems,” IEEE Netw., vol. 34, no. 3, pp. 134–142, 2020.
[4] E. C. Strinati, S. Barbarossa, J. L. Gonzalez-Jimenez, et al., “6G: the next frontier: From holographic messaging to artificial
intelligence using subterahertz and visible light communication,” IEEE Vehicular Technology Magazine, vol. 14, no. 3, pp. 42–
58, 2019.
[5] L. Imoize, O. Adedeji, N. Tandiya, and S. Shetty, “6G enabled smart infrastructure for sustainable society: Opportunities,
challenges, and research roadmap,” Sensors, vol. 21, no. 5, pp. 1–57, Mar. 01, 2021.
[6] Y. Zhao, G. Yu, and H. Xu, “6G mobile communication networks: Vision, challenges, and key technologies,” Scientia Sinica
Information is, vol. 49, no. 8, pp. 963–987, 2019.
[7] Y. L. Lee, D. Qin, L. C. Wang, and G. H. Sim, “6G massive radio access networks: Key applications, requirements and
challenges,” IEEE Open Journal of Vehicular Technology, vol. 2, pp. 54–66, 2021.
[8] T. S. Rappaport, Y. Xing, O. Kanhere, et al., “Wireless communications and applications above 100 GHz: Opportunities and
challenges for 6G and beyond,” IEEE Access, vol. 7, pp. 78729–78757, 2019.
[9] M. Ylianttila, R. Kantola, A. Gurtov, L. Mucchi, I. Oppermann, Z. Yan,T. H. Nguyen, F. Liu, T. Hewa, M. Liyanage et al., “6G
White Paper: Research Challenges for Trust, Security and Privacy,” arXiv preprint arXiv:2004.11665, 2020.
[10] I. Ahmad, T. Kumar, M. Liyanage, J. Okwuibe, M. Ylianttila, and A. Gurtov, “Overview of 5G Security Challenges and
Solutions,” IEEE Communications Standards Magazine, vol. 2, no. 1, pp. 36–43, 2018.
[11] R. Khan, P. Kumar, D. N. K. Jayakody, and M. Liyanage, “A survey on security and privacy of 5g technologies: Potential
solutions, recent advancements, and future directions,” IEEE Communications Surveys & Tutorials, vol. 22, no. 1, pp. 196–248,
2019.
[12] P. Ranaweera, A. D. Jurcut, and M. Liyanage, “Survey on Multi-Access Edge Computing Security and Privacy,” IEEE
Communications Surveys Tutorials, pp. 1–1, 2021.
[13] S. Wijethilaka and M. Liyanage, “Survey on network slicing for internet of things realization in 5g networks,” IEEE
Communications Surveys & Tutorials, 2021
[14] N. Wang, P. Wang, A. Alipour-Fanid, L. Jiao, and K. Zeng, “Physical- Layer Security of 5G Wireless Networks for IoT:
Challenges and Opportunities,” IEEE Internet of Things Journal, vol. 6, no. 5, pp. 8169– 8181, 2019.
[15] Y. Yang, L. Wu, G. Yin, L. Li, and H. Zhao, “A Survey on Security and Privacy Issues in Internet-of-Things,” IEEE Internet of
Things Journal, vol. 4, no. 5, pp. 1250–1258, 2017.
[16] C. Benzaid and T. Taleb, “AI for Beyond 5G Networks: A Cyber- Security Defense or Offense Enabler?” IEEE Network, vol. 34,
no. 6, pp. 140–147, 2020.
[17] T. Hewa, G. Gür, A. Kalla, M. Ylianttila, A. Bracken, and M. Liyanage, “The Role of Blockchain in 6G: Challenges,
Opportunities and Research Directions,” in 2020 2nd 6G Wireless Summit (6G SUMMIT). IEEE, 2020, pp. 1–5,.
[18] A. Dogra, R. K. Jha, and S. Jain, “A Survey on beyond 5G Network with the Advent of 6G: Architecture and Emerging
Technologies,” IEEE Access, 2020.
[19] G. Blinowski, “Security of Visible Light Communication Systems — A Survey,” Physical Communication, vol. 34, pp. 246–260,
2019.
[20] W. Saad, M. Bennis, and M. Chen, “A vision of 6G wireless systems: Applications, trends, technologies, and open research
problems,” IEEE Netw., vol. 34, no. 3, pp. 134–142, 2020.
[21] S. R. A. Jafri, T. Hamid, R. Mahmood, M. A. Alam, T. Rafi, M.Z. Ul Haque, and M. W. Munir, “Wireless brain computer
interface for smart home and medical system,” Wireless Personal Communications, vol. 106, no. 4, pp. 2163–2177, 2019.
[22] P. K. Padhi and F. Charrua-Santos, “6G enabled industrial internet of everything: Towards a theoretical framework,” Applied
System Innovation, vol. 4, no. 1, p. 11, Feb. 2021.
[23] I. H. Abdulqadder, S. Zhou, D. Zou, I. T. Aziz, and S. M. A. Akber, “Multi-layered Intrusion Detection and Prevention in the
SDN/NFV enabled Cloud of 5G Networks using AI-based Defence Mechanisms,” Computer Networks, vol. 179, p. 107364,
2020.
[24] R. Santos, D. Souza, W. Santo, A. Ribeiro, and E. Moreno, “Machine Learning Algorithms to detect DDoS Attacks in SDN,”
Concurrency and Computation: Practice and Experience, vol. 32, no. 16, p. e5402, 2020.
[25] M. Zolanvari, M. A. Teixeira, L. Gupta, K. M. Khan, and R. Jain, “Machine Learning-Based Network Vulnerability Analysis of
Industrial Internet of Things,” IEEE Internet of Things Journal, vol. 6, no. 4, pp. 6822–6834, 2019.
[26] H. Fang, X. Wang, and S. Tomasin, “Machine Learning for Intelligent Authentication in 5G and Beyond Wireless Networks,”
IEEE Wireless Communications, vol. 26, no. 5, pp. 55–61, 2019.
[27] J. Biamonte, P. Wittek, N. Pancotti, P. Rebentrost, N. Wiebe, andS. Lloyd, “Quantum Machine Learning,” Nature, vol. 549, no.
7671, pp. 195–202, 2017.
[28] L. Xiao, G. Sheng, S. Liu, H. Dai, M. Peng, and J. Song, “Deep Reinforcement Learning-enabled Secure Visible Light
communication against Eavesdropping,” IEEE Transactions on Communications, vol. 67, no. 10, pp. 6994–7005, 2019.
[29] M. Liyanage, J. Salo, A. Braeken, T. Kumar, S. Seneviratne, and M. Ylianttila, “5G Privacy: Scenarios and Solutions,” in 2018
IEEE 5G World Forum (5GWF), 2018, pp. 197–203.
[30] Y. Sun, J. Liu, J. Wang, Y. Cao, and N. Kato, “When Machine Learning meets Privacy in 6G: A Survey,” IEEE Communications
Surveys & Tutorials, vol. 22, no. 4, pp. 2694–2724, 2020.
[31] J. Li, X. Kuang, S. Lin, X. Ma, and Y. Tang, “Privacy Preservation for Machine Learning Training and Classification based on
Homomorphic Encryption Schemes,” Information Sciences, vol. 526, pp. 166–179, 2020.
[32] T. Zhang, T. Zhu, P. Xiong, H. Huo, Z. Tari, and W. Zhou, “Correlated Differential Privacy: Feature Selection in Machine
Learning,” IEEE Transactions on Industrial Informatics, vol. 16, no. 3, pp. 2115–2124, 2020.
[33] J.-H. Cho, D. P. Sharma, H. Alavizadeh, S. Yoon, N. Ben-Asher, T. J. Moore, D. S. Kim, H. Lim, and F. F. Nelson, “Toward
Proactive, Adaptive Defence: A Survey on Moving Target Defense,” IEEE Com- munications Surveys & Tutorials, vol. 22, no.
1, pp. 709–745, 2020.