artificial intelligence and Cybersecurity of the intern

Question

artificial intelligence and Cybersecurity of the internet of things

MANZER KHAN[T] · Accepted Answer

Research Links
1.	Abdullahi, M., Baashar, Y., Alhussian, H., Alwadain, A., Aziz, N., Capretz, L. F., & Abdulkadir, S. J. (2022). Detecting Cybersecurity Attacks in Internet of Things Using Artificial Intelligence Methods: A Systematic Literature Review. Electronics, 11(2), 198. https://doi.org/10.3390/electronics11020198
2.	Kuzlu, M., Fair, C., & Guler, O. (2021). Role of artificial intelligence in the Internet of Things (IoT) cybersecurity. Discover Internet of things, 1(1), 1-14. https://doi.org/10.1007/s43926-020-00001-4 
3.	Shafiq, M., Tian, Z., Sun, Y., Du, X., & Guizani, M. (2020). Selection of effective machine learning algorithm and Bot-IoT attacks traffic identification for internet of things in smart city. Future Generation Computer Systems, 107, 433-442. https://doi.org/10.1016/j.future.2020.02.017 
4.	Almiani, M., AbuGhazleh, A., Al-Rahayfeh, A., Atiewi, S., & Razaque, A. (2020). Deep recurrent neural network for IoT intrusion detection system. Simulation Modelling Practice and Theory, 101, 102031. https://doi.org/10.1016/j.simpat.2019.102031

Justification
Technology like the Internet of Things (IoT) is now widely used in everything from classical equipment to everyday domestic products, dramatically enhancing our standard of living.  However, the spread nature, sheer number of items, and openness of IoT systems have made them a prime target for cyber attackers. One of the first steps in securing IoT networks is to identify intruders. Artificial intelligence (AI) can be used to spot unusual patterns of behavior on IoT devices, indicating an attack.’

Abstract
The concept of the Internet of Things (IoT) has brought about improvements in terms of accessibility, authenticity, availability, scalability, privacy, and interoperability. However, the IoT is susceptible to attacks because it has a large number of attack surfaces. As a result of this, there is a significant amount of interest in the topic of cybersecurity for the Internet of Things. The application of artificial intelligence (AI) is useful to identify aberrant patterns of behaviour that would indicate an attack on the internet of things devices. As a result, the purpose of this study is to determine whether or not AI is useful in reducing the risk of cybersecurity threats posed by the internet of things (IoT). This was accomplished by conducting a comprehensive literature analysis with a particular emphasis on more recent developments.

Introduction
Since roughly 2008, when the Internet of Things (IoT) was first introduced, its popularity is skyrocketing, and it is now commonplace in many households and companies (Evans 2011). Since its invention, the Internet of Things (IoT) has been difficult to define, but it can be best defined as a network of digital and analog machines and computing devices with unique identifiers (UIDs) that can communicate data without human interaction (Kujlu et al., 2021). In terms of device connectivity, the Internet of Things (IoT) idea has improved accessibility, authenticity, availability, scalability, secrecy, and interoperability (Lu & Da 2018). IoT, on the other hand, is vulnerable to assaults because of its numerous attack surfaces and lack of safety standardizations and criteria. When it comes to hacking into IoT systems, attackers have a wide range of options. Because of this, there is a lot of interest in cybersecurity for the Internet of Things. The use of Artificial Intelligence (AI) to detect anomalous behavior that may suggest an assault on IoT systems is one example of this (Lakhani 2018). It is easier for cybercriminals to exploit a single IoT device than it is for security specialists to guard against several threats. Because of this, cybercriminals are increasingly turning to artificial intelligence (AI) as a means of evading complex algorithms that would otherwise detect suspicious activities (Kujlu et al., 2021). When it comes to the Internet of Things (IoT), AI has attracted a lot of attention. Cybersecurity applications for the Internet of Things (IoT) have seen a rise in the use of artificial intelligence (AI) techniques, such as decision trees and linear regression. 
Market analysts expect artificial intelligence in cybersecurity to reach USD 38.2 billion by 2026, up 23.3% from USD 8.8 billion in 2019. Growing connectivity, rising cyber threats, concerns about data privacy, and the growing frailty of Wi-Fi networks are all major factors in the market's growth (Market Research Firm, n.d.). 
In this regard, a detailed literature study on the implications of AI in improving the cybersecurity of the internet of things is presented and discussed in this article. Also included are some recommendations for the future.  
Literature Review
Because of the enormous volume of data that must be processed, IoT networks must now implement security measures. A growing number of cyberattacks are targeting these systems, and their sophistication is increasing. Therefore, for a machine learning neural network method, Lavirra-Novo et al (2021) offered many data augmentation based on traffic categorization. There are two benchmark datasets utilized in this study: UGR16 and the UNSW-NB15, as well as the most commonly used dataset, KDD99. Scalar and normalizing functions were used to assess the preprocessing methods. There were four categories of features used to classify the preprocessing models: connection attributes, content characteristics, statistical characteristics, and a group made up of traffic-based functionalities and connection direction-based traffic characteristics. Each of these groups was used to apply the preprocessing models. Data preprocessing techniques will be used to evaluate this categorization to come up with the best accurate model. Up to 45 percent, more accuracy can be achieved by utilizing network traffic categorization and a variety of preprocessing techniques. As a result, machine learning algorithms are able to correctly classify factors that are associated with probable assaults more accurately when they are processed prior. Due to the time and complexity required to train the models, this technique has a number of drawbacks.
Two-level ensemble attack detection for industrial control systems has been proposed by Jahromi et al. (2021). A decision tree and a unique ensemble deep representation-learning model are used to detect attacks. Moreover, an ensemble deep neural network was developed to guide the attribution of an attack. Using real-world data from the pipeline project and water treatment systems, the proposed model is tested. Research shows that the conceptual model outperforms competing alternatives of the same computing complexity. However, the key drawback is that the precision of the detection needs to be improved.
An IoT feature extraction as well as intrusion detection technique for intelligent cities has been presented by Li et al (2019) using a deep migration learning model, that incorporates intrusion detection technology. This work introduces the migration learning model modeling scheme with data feature extraction based on current literature and techniques. A KDD CUP 99 dataset was utilized as the exploratory data set, with 10% of the dataset being used as training examples. The suggested approach was also compared to current algorithms. It has been demonstrated in the experiments that the proposed technique is more efficient and less time-consuming to detect. However, the key drawback is that the precision of the detection needs to be improved.
An artificial comprehensive monitoring system for Fog protection against cyber-attacks was presented by Almiani et al (2020). Multi-layered recurrent neural networks are used in the proposed model for Fog computing security that is near the end-users and IoT devices. NSL-KDD, a well-balanced variant of the difficult dataset, was used to show our model. As the experimental and simulation findings confirmed the stability and robustness of the proposed model in terms of a variety of performance metrics, we included two more metrics: Mathew correlation and Cohen's Kappa coefficients for deeper insight. Thus, this model has a greater level of robustness and is better at predicting time series. However, this model's drawback is the identification of the requirement for improved precision as well as the explosion of gradients.
Using machine learning techniques (LR, SVM, DT, RF, ANN, and KNN), Rashid et al. (2020) investigated an assault and anomaly detection technique in a smart city to combat and alleviate IoT cybersecurity threats. For example, instead of relying just on a single classifier, the researchers looked into ensemble methods like bagging, boosting, and stacking. The research also took into account a method that hasn't been extensively explored before: the convergence of feature selection, cross-validation, and multi-class classification. Results from a recent cyber attack dataset showed that the suggested technique successfully identified assaults and the stacking ensemble model beats comparable models in terms of accuracies, precision, recall, and F1-Score. A major drawback of this approach is the time and complexity required to train the model and decipher the results; hence, more work must be done on these aspects before it can be considered fully functional.
In the modern intelligent power grid, electricity generation and consumption are efficiently managed while coping with a wide range of security risks. Disruptions to the electrical grid can be caused by both spontaneous and man-made occurrences. As a result, operators must be able to recognize the precise sorts of power system disturbances and the causes that generate them in order to respond effectively. With this in mind, Wang et al. (2019) developed a machine learning-based attack detection model for the power system that can be trained utilizing logs and other data from phasor measuring devices (PMUs). The data is sent to various machine learning models, using the random forest as the primary classifier in AdaBoost, before the researchers feature the construction engineering. An open-source simulated power network data set of 37 event scenarios were used to test the model. In the end, the researchers analyzed the suggested model to other models using various evaluation measures. This model outperformed eight recently developed algorithms in terms of accuracy and detection rate, according to the findings of the experiments. It is, however, limited by the fact that the interpretation of results needs to be improved, as well as the time and complexity of model training and computing.
An anomaly detection approach described by Bhatia et al (2019) was based on network-centric, behavior-learning-based anomaly identification. Unsupervised machine learning was used to detect distinct sorts of DDoS attacks in real-time using the dependability of TCP traffic from IoT devices. The suggested ML classifier is able to distinguish between normal and abnormal traffic based on a minimal number of variables. SDN-based procedures for filtering attack traffic near the source can be implemented using this methodology in a larger system to identify compromised endpoints despite IP spoofing. For new and unknown assaults, this unsupervised ML methodology is easier to implement and more accurate than supervised ones. However, the biggest drawback is the need for improvement in the interpretation of results.
To address the cybersecurity-related issue, Shafiq et al. (2020) developed an entirely new conceptual framework and a hybrid algorithm. Machine learning algorithms were first trained on a BoT-IoT identification dataset, which had 44 effective features selected from a large number of attributes. For malicious and anomalous traffic identification, five successful machine learning algorithms were selected, and the most extensively used ML algorithm performance review criteria were also picked. It was determined that a bijective soft set technique and its algorithm might be used to identify IoT anomalies and intrusion traffic. To test their new technique, the researchers used a soft set approach that is based on bijective soft sets. This algorithmic model was shown to be a useful tool for narrowing down the field of potential machine learning algorithms. Its two primary drawbacks are that it requires a lot of time to train and that it is quite computationally demanding.
However, Comiter 2019 reported that the use of malevolent AI like fuzzing and symbolic execution to assist attacks, often to thwart intrusion detection algorithms in the IoT, or to attack advantageous AI in such a way that the AI works against its system, is becoming increasingly common.
Conclusion and recommendation
To safeguard IoT systems, this article seeks to give researchers and cybersecurity experts a valuable tool for studying IoT in the context of cybersecurity and AI. An additional focus will be on future developments in technology and how they will interact with one another. 
As the use of IoT devices grows in numerous areas, including IIoT, healthcare IoT, energy IoT, and CPS, we encourage further exploration of the current AI algorithms by proposing next-generation techniques in IoT confidentiality and protection.
A variety of artificial intelligence (AI) approaches based on ML/DL algorithms, including hybrid approaches that mix multiple algorithms have been suggested and implemented by academics. ML/DL approaches with improved performance can be introduced by upgrading existing intelligent architectural frameworks.
Our findings also confirm the existence of many forms of cybersecurity assaults in the IoT context. However, researchers can still look for flaws in other aspects of the attack as well. Additionally, the most susceptible layer in IoT must be recognized.
Improved detection efficiency, as well as a focus on recent real-world datasets, are needed to detect the latest dangers from the Internet of Things. These findings may inspire new AI models to protect the Internet of Things (IoT) against cybersecurity threats and attacks, according to our research.
Gantt Chart

References
Almiani, M., AbuGhazleh, A., Al-Rahayfeh, A., Atiewi, S., & Razaque, A. (2020). Deep recurrent neural network for IoT intrusion detection system. Simulation Modelling Practice and Theory, 101, 102031.
 Artificial Intelligence in cybersecurity market. Market Research Firm. (n.d.). Retrieved June 3, 2022, from https://www.marketsandmarkets.com/Market-Reports/artificial-intelligence-security-market-220634996.html  
Bhatia, R., Benno, S., Esteban, J., Lakshman, T. V., & Grogan, J. (2019, December). Unsupervised machine learning for network-centric anomaly detection in IoT. In Proceedings of the 3rd acm conext workshop on big data, machine learning and artificial intelligence for data communication networks (pp. 42-48).
Comiter, M. (2019). Attacking artificial intelligence. Belfer Center Paper, 2019-08.
Evans, D. (2011). The Internet of Things—How the Next Evolution of the Internet Is Changing Everything, White Paper. Cisco Internet Business Solutions Group (IBSG).
Jahromi, A. N., Karimipour, H., Dehghantanha, A., & Choo, K. K. R. (2021). Toward Detection and Attribution of Cyber-Attacks in IoT-Enabled Cyber–Physical Systems. IEEE Internet of Things Journal, 8(17), 13712-13722.
Kuzlu, M., Fair, C., & Guler, O. (2021). Role of artificial intelligence in the Internet of Things (IoT) cybersecurity. Discover Internet of things, 1(1), 1-14.
Lakhani, A. (2018). The role of artificial intelligence in IoT and OT security.
Larriva-Novo, X., Villagrá, V. A., Vega-Barbas, M., Rivera, D., & Sanz Rodrigo, M. (2021). An IoT-focused intrusion detection system approach based on preprocessing characterization for cybersecurity datasets. Sensors, 21(2), 656.
Li, D., Deng, L., Lee, M., & Wang, H. (2019). IoT data feature extraction and intrusion detection system for smart cities based on deep migration learning. International journal of information management, 49, 533-545.
Lu, Y., & Da Xu, L. (2018). Internet of Things (IoT) cybersecurity research: A review of current research topics. IEEE Internet of Things Journal, 6(2), 2103-2115.
Rashid, M. M., Kamruzzaman, J., Hassan, M. M., Imam, T., & Gordon, S. (2020). Cyberattacks detection in iot-based smart city applications using machine learning techniques. International Journal of environmental research and public health, 17(24), 9347.
Shafiq, M., Tian, Z., Sun, Y., Du, X., & Guizani, M. (2020). Selection of effective machine learning algorithm and Bot-IoT attacks traffic identification for internet of things in smart city. Future Generation Computer Systems, 107, 433-442.
Wang, D., Wang, X., Zhang, Y., & Jin, L. (2019). Detection of power grid disturbances and cyber-attacks based on machine learning. Journal of information security and applications, 46, 42-52.

artificial intelligence and Cybersecurity of the internet of things

Get Instant Homework HelpOn Your Mobile

Get Instant Homework HelpOn Your Mobile

Get Instant Homework HelpOn Your Mobile