Machine Learning and Feature Selection Based Ransomware Detection Using Hexacodes

Evaluating Large Language Models in Ransomware Negotiation: A Comparative Analysis of ChatGPT and Claude

“…The field of ransomware detection has seen significant advancements, with a variety of methodologies being developed and refined [26], [27]. Initially, the focus was predominantly on signature-based detection, a method that relies on identifying known ransomware signatures [3].…”

Using ChatGPT to Analyze Ransomware Messages and to Predict Ransomware Threats

“…This segment of the discussion centers around the incorporation of artificial intelligence in the realm of ransomware threat identification, with a special focus on identifying zero-day ransomware assaults and emerging variants, which involves delving into the development of models based on deep learning for the detection of ransomware, paying special attention to unsupervised methods of extracting features and the application of ensemble classifiers that are sensitive to the cost implications of false classifications [17], [5]. These models' proficiency in countering zero-day incursions, along with their capacity to expedite the identification of such threats, is a topic of particular interest [21], [7].…”

“…A key area of discussion is the role AI plays in refining the process of victim selection, enabling attackers to tailor their strategies to target individuals or organizations with precision [23], [17]. The capacity of AI to customize attack payloads for each victim, thereby enhancing the efficacy of the assault, is a significant aspect of this exploration [13].…”

Crypto-ransomware detection using machine learning models in file-sharing network scenarios with encrypted traffic

“…Recent surveys (Berrueta et al, 2019;Bijitha et al, 2020;Herrera Silva et al, 2019) have analysed and compared a total of more than 70 different crypto-ransomware detection tools. Ransomware detection tools have traditionally used similar techniques to antivirus tools, such as those based on the static analysis of program binaries before they are run (Hasan & Rahman, 2017;Reddy et al, 2021;Shaukat & Ribeiro, 2018). These techniques offer prevention, however, they are prone to false negatives when new binaries are encountered or obfuscation techniques are used (Chen et al, 2017;Vidyarthi et al, 2019), they cannot cope with the large number of signatures created by the new ransomware-as-a-service offerings (Nieuwenhuizen, 2016), and they are oblivious to file-less ransomware variants (Victor, 2020).…”

“…In some cases it only establishes thresholds to the measured metrics (Kharraz et al, 2016;Mbol et al, 2016;Paik et al, 2016), while most often a combined metric from a large number of indicators is built (Moore, 2016;Morato et al, 2018;Scaife et al, 2016). In recent years, machine learning techniques have gained popularity thanks to their ease of use and their capacity of searching for patterns in a large number of features (Ahmadian & Shahriari, 2016;Ahmed et al, 2020;Almashhadani et al, 2019;Arabo et al, 2020;Continella et al, 2016;Hasan & Rahman, 2017;Hirano & Kobayashi, 2019;Lee et al, 2019;Lu et al, 2017;Moussaileb et al, 2018;Reddy et al, 2021;Roy & Chen, 2020;Scaife et al, 2016;Sgandurra et al, 2016;Vinayakumar et al, 2017).…”