Review of Artificial Intelligence Cyber Threat Assessment Techniques for Increased System Survivability

Doukas, Nikolaos; Stavroulakis, Peter; Bardis, Nikolaos G.

doi:10.1007/978-3-030-62582-5_7

Cited by 6 publications

(6 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Following the rules and being ethical is important to stop people from using things in the wrong way and causing harm Doukas et.al. [16]. People are worried that AI systems might make existing biases and discrimination worse.…”

Section: Challenges and Limitationsmentioning

confidence: 99%

“…This detailed review will help with future studies and advancements in the area of IoT security Doukas et.al. [16] research looks at how artificial intelligence (AI) can help make systems better at surviving cyber attacks. AI helps systems find and react to threats faster and better without human help.…”

Section: Related Workmentioning

confidence: 99%

“…Emerging AI and Big Data technologies are changing how businesses handle risk and protect against online threats Doukas et.al. [16]. Edge computing can help make decisions and assess risks in real-time in IoT and industrial settings.…”

Section: Emerging Trends and Technologiesmentioning

confidence: 99%

See 2 more Smart Citations

A Systematic Literature Review: AI, DL and Machine Learning inCyber Risk Management

Hamid,

Rahman

2024

Preprint

View full text Add to dashboard Cite

The integration of AI and ML in both climate change and cyber risk management has significantly bolstered their effectiveness. These advanced technologies enable organizations to handle extensive datasets, whether related to climate patterns or cyber activities, with greater efficiency and making more informed decisions. AI and machine learning are utilized in cyber security to predict potential threats, automate tasks, and manage cyber security risks. Ensuring the accuracy of information, comprehending the functionality of the model, and safeguarding individuals' data pose challenges. It is very important to use personal information ethically and protect it. Researchers should prioritize accuracy in the information utilized, clarity in AI models, and fairness in AI decision-making in the future. New technologies like edge computing, federated learning, explainable AI, and quantum computing are changing the way we manage risk and keep our digital information safe. These new advancements provide opportunities to gain a better understanding of risks, identify threats more rapidly, and improve decision-making. Still, there are challenges concerning data accuracy, understanding model operations, and privacy issues that must be resolved before these technologies can effectively address cyber risks.

show abstract

Section: Challenges and Limitationsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

A Systematic Literature Review: AI, DL and Machine Learning inCyber Risk Management

Hamid,

Rahman

2024

Preprint

View full text Add to dashboard Cite

show abstract

“…In the rapidly evolving landscape of cybersecurity, malware poses a significant threat to information security, with new variants emerging at an alarming rate [1,2,3,4]. The sophistication of these malicious programs has escalated, particularly in the context of evasive malware, which is designed to thwart detection mechanisms employed by traditional antivirus and cybersecurity measures [5,6,7,8,9,10].…”

Section: Introductionmentioning

confidence: 99%

“…The sophistication of these malicious programs has escalated, particularly in the context of evasive malware, which is designed to thwart detection mechanisms employed by traditional antivirus and cybersecurity measures [5,6,7,8,9,10]. The detection and analysis of such evasive malware remain a constant challenge in the field, necessitating advanced methods and innovative approaches [2,6]. Traditional methods, primarily based on signature detection, are increasingly insufficient as they struggle to cope with the adaptive and polymorphic nature of modern malware, which signals the urgent need for more dynamic and intelligent analysis methods capable of identifying and mitigating these sophisticated threats [6,1].…”

Section: Introductionmentioning

confidence: 99%

A Novel Approach of Evasive Malware Analysis Through Binary Opcode and BERT

Wang,

2024

Preprint

View full text Add to dashboard Cite

This study addressed the growing challenge of evasive malware detection by introducing an innovative approach that integrates binary opcode analysis with the advanced machine learning capabilities of BERT (Bidirectional Encoder Representations from Transformers). Focusing on the rapidly evolving landscape of cybersecurity, this study explored the intricacies of malware behavior, particularly in the context of opcode obfuscation and direct OS lower-level API access. Through a comprehensive methodology, the study demonstrated the effectiveness of combining traditional disassembly techniques with state-of-the-art deep learning algorithms. The findings revealed a significant distinction between malware and benign software in terms of their opcode structures and system interactions. This study not only offered a novel perspective in malware analysis, but also set the stage for future advancements in cybersecurity tools and techniques. The integration of binary opcode analysis and BERT represented a vital development in cybersecurity, providing enhanced detection capabilities and a deeper understanding of sophisticated malware threats.

show abstract