Enhancing Ransomware Attack Detection on Cloud-Encrypted Data

Version	Summary	Created by	Modification	Content Size	Created at	Operation
1		Grzegorz Nowakowski	--	3467	2023-12-09 12:38:52	\|
2	format correct	Catherine Yang	-16 word(s)	3451	2023-12-11 02:10:44	\|

This entry is adapted from the peer-reviewed paper 10.3390/electronics12183899

Ransomware attacks on cloud-encrypted data pose a significant risk to the security and privacy of cloud-based businesses and their consumers. RANSOMNET+, a state-of-the-art hybrid model that combines Convolutional Neural Networks (CNNs) with pre-trained transformers, to efficiently take on the challenging issue of ransomware attack classification. RANSOMNET+ excels over other models because it combines the greatest features of both architectures, allowing it to capture hierarchical features and local patterns. The findings demonstrate the exceptional capabilities of RANSOMNET+. The model had a fantastic precision of 99.5%, recall of 98.5%, and F1 score of 97.64%, and attained a training accuracy of 99.6% and a testing accuracy of 99.1%. The loss values for RANSOMNET+ were impressively low, ranging from 0.0003 to 0.0035 throughout training and testing. RANSOMNET+ excelled over the other two models in terms of F1 score, accuracy, precision, and recall. The algorithm’s decision-making process was also illuminated by RANSOMNET+’s interpretability analysis and graphical representations. The model’s openness and usefulness were improved by the incorporation of feature distributions, outlier detection, and feature importance analysis. Finally, RANSOMNET+ is a huge improvement in cloud safety and ransomware research. As a result of its unrivaled accuracy and resilience, it provides a formidable line of defense against ransomware attacks on cloud-encrypted data, keeping sensitive information secure and ensuring the reliability of cloud-stored data. Cybersecurity professionals and cloud service providers now have a reliable tool to combat ransomware threats thanks to this research.

ransomware attack detection transfer learning deep learning ensemble models cloud-encrypted data cybersecurity

1. Introduction

Due to the growth of cloud services and the ever-increasing volume of online data, safeguarding the privacy and accessibility of encrypted data stored in the cloud has become an essential problem ^[1]. Cyber threats such as ransomware attacks have become increasingly common and destructive in recent years: “malicious software that sneaks onto computers, encrypts important files, and then demands a ransom to decrypt them” is described in ^[2]. These attacks on individuals, organizations, and even government bodies have resulted in major financial losses and data breaches. The two most prominent methods for detecting ransomware—signature-based methods and rule-based systems—have been unable to keep up with the rapidly evolving nature of ransomware attacks ^[3]. Updating signatures and rules often to detect new variants might decrease reaction times and leave systems vulnerable to new assaults. The encrypted nature of cloud data limits accessibility and analysis, making it difficult to detect ransomware assaults using conventional approaches ^[3]^[4].

Transfer learning is a technique that allows the transfer of knowledge learned from pre-trained models on large-scale datasets, which can be used to improve detection accuracy even in the absence of labeled ransomware samples. Using transfer learning, the suggested strategy compensates for the scarcity of ransomware data and enhances detection capabilities ^[4]^[5].

The devastating financial losses and compromises in data privacy and security that can arise from ransomware attacks have made them a prominent cybersecurity worry in recent years. Ransomware is malicious software that infiltrates systems and encrypts data in order to demand payment. These attacks have had devastating results for individuals, businesses, and government institutions. Cloud-based systems that contain huge amounts of sensitive data have been a common target for ransomware attacks ^[5]^[6] due to their broad use and potential for large-scale effects. Since ransomware is constantly adapting, signature-based approaches and rule-based systems for detecting it have proven insufficient. Updating signatures and rules often to detect new variants might decrease reaction times and leave systems vulnerable to new assaults. As a result of the encrypted nature of cloud data, traditional methods of detecting ransomware assaults ^[7] are ineffective due to a lack of data accessibility and analysis.

This inquiry was prompted by the increasing frequency and severity of ransomware attacks that target cloud environments. Traditional ransomware detection methods are useless because ransomware strains constantly evolve. The encrypted nature of cloud data presents additional challenges for detection systems, making it all the more important to explore innovative ways that may effectively identify ransomware attacks in this encrypted context ^[8]. The employment of transfer learning and deep learning ensembles is motivated by the expectation of enhanced detection precision and adaptability. Using deep learning ensembles in detection systems can improve cloud-based security by capturing more features ^[9]. Using cutting-edge methods such as transfer learning and deep learning ensembles for ransomware detection ^[10], this research aims to strengthen the protections afforded by cloud-based systems and shield sensitive data from dangerous cyber-attacks.

The development of a deep learning ensemble framework for ransomware detection is another important step forward. Using a large number of deep learning models to collect diverse characteristics, the ensemble framework improves detection performance and robustness. This new technique can help in the design of better real-time detection systems for ransomware ^[11].

2. Enhancing Ransomware Attack Detection Using Transfer Learning and Deep Learning Ensemble Models on Cloud-Encrypted Data

It is now a major security risk for computers to become infected with ransomware, which encrypts user data. Modern ransomware strains are more dangerous than their older counterparts because they employ complex obfuscation tactics and can operate without an active C2 server. The need for a unified framework to aid with ransomware detection, prevention, and mitigation is the driving force for this research. The article ^[12] introduced the Detection Avoidance Mitigation (DAM) framework, which organizes and describes existing methods of countering ransomware. By conducting a comprehensive review and synthesis of the relevant literature, similarly in ^[13] accomplished an important advancement in the field by establishing a uniform methodology. The Djvu Ransomware case study is intended to demonstrate the ransomware’s tactics and to provide countermeasures. DAM is a framework developed to fortify systems against ransomware.

There is a major risk to network infrastructure because of the rise of hostile threats against computer networks and digital services. In order for one website to link to another, a system called the “domain name system” (DNS) must be in place. The difficulties of discovering hidden tunnels and avoiding conventional detection methods must be surmounted if these DNS intrusions ^[14] are to be spotted. Statistical analysis and Bi-directional Recurrent Neural Network (BRNN) methods are used to construct an intrusion detection model in this research paper, with the goal of revealing hostile DNS over HTTPS (DoH) requests made through covert channels. The method is 100% accurate in identifying malicious DoH searches using data from the Canadian Institute for Cybersecurity’s CIRA-CIC-DoHBrw-2020 dataset. The proposed model, which uses fewer features than competing methods, yields better results at a higher throughput during training and testing.

Customers with concerns about malware and data loss in the cloud are not crazy. Although the effectiveness of virus detection has been the subject of multiple research, renters’ rights to privacy in the cloud are rarely taken into account. In this research work ^[6] introduced a novel cloud-based malware detection technique based on semi-supervised transfer learning (SSTL). Detection, foresight, and transfer are the pillars upon which the model rests. Researchers have developed a byte classifier based on recurrent neural networks to safeguard user data stored in the public cloud. There is not enough data for the byte classifier to improve its supervised learning performance beyond 94.72%. Accuracy in the prediction phase has been improved to 99.69% thanks to a new ASM classifier. To enhance the byte classifier’s training, the transfer module uses semi-supervised learning to merge predicted labels and byte attributes from an unlabeled dataset. The accuracy of the detection part was improved using semi-supervised transfer learning, and testing on Kaggle malware datasets showed an increase from 94.72% to 96.9%. This method improves malware detection accuracy and addresses tenants’ privacy concerns.

Recent advancements in machine learning and deep learning for ransomware detection are the topic of this state-of-the-art review ^[7]. The critical need to defend computer systems from ransomware attacks sparked this investigation. The ability of machine learning and deep learning techniques to identify zero-day attacks and build prediction models based on ransomware behavior has increased their widespread adoption. The author of this review included research using machine learning or deep learning methodologies for ransomware detection because those methods have received a high number of citations. The author also ran studies to determine how malware evolution might have affected the results. The author also speculates on the possible directions that ransomware could go in the future, including spreading to IoT devices and seeing increased use in both households and businesses.

There has been a rise in the use of computers and the Internet for the storage and transfer of private information. The term “cyber-ransomware” elicits broad feelings of fear and alarm. Crypto-ransomware is a type of malicious software that encrypts data and then demands payment to decrypt it. Numerous machine learning-based detection investigations have been conducted; however, cybercriminals are always developing new forms of encryption to avoid detection. Users’ private information may still be at risk if ransomware attacks go undetected. Using your phone’s camera to capture events, as suggested in ^[15], is advantageous because it negates the need for a backup. SVM analyzes 22 encrypted file formats, extracts unique features from each, and achieves an 85.17 percent detection rate. This method’s high performance and efficacy are demonstrated by the fact that its detection rate is greater than 92% when combined with the SVM kernel Trick (Poly).

New exploits in computer systems are discovered and used by ransomware as they evolve. Researchers and practitioners alike rely on machine learning techniques for ransomware detection and mitigation, particularly transfer learning. Transfer learning, in which models are used to solve other problems, can help make ransomware detection systems more accurate and adaptable. There is a growing corpus of work on ransomware detection, and it is getting increasingly difficult to identify the machine learning algorithms and transfer learning techniques used in this research. The goal of this study ^[16] was to help researchers better understand ransomware detection frameworks and common machine learning algorithms, particularly those that use transfer learning to extract ransomware’s dynamic properties. These ransomware detection frameworks, datasets, and issues are investigated in detail. Researchers and practitioners can use the results of this comparison study to improve their usage of transfer learning to detect ransomware.

Because of its capacity to encrypt data and block access to it, ransomware poses a significant security concern to enterprises. The article ^[12] developed a technique for analyzing traffic from file-sharing platforms to improve the detection of ransomware assaults. Deep learning (DL) and transfer learning (TL) are two examples of machine learning algorithms used to monitor client-server communication for suspicious behavior indicative of ransomware during file reading and overwriting. Both clear-text and encrypted file-sharing techniques work with the solution without any modifications needed. After comparing multiple machine learning models, the most successful one is chosen for validation. These results show that the author’s taught and tested detection model is successful against all ransomware binaries, even those that have not yet been found. ‘Not infected’ traffic from actual users was used alongside more than 70 ransomware files representing 26 different strains during the model’s training and testing phases. To guarantee the procedure is reliable, this study investigates the number of false positives and the amount of file encryption used before detection. The plan uses DL, TL, and traffic analysis to improve ransomware detection, safeguarding the organization’s infrastructure and data in the process.

Ransomware becomes a major issue for everyone from individuals to large corporations when files can be accessed from various servers. The research study in ^[12] proposed a method to combat this problem by keeping an eye on file-sharing traffic for signs of crypto-ransomware. The program detects ransomware-like behavior in file reading and overwriting by keeping tabs on client-server communication and using machine learning techniques. Because it may be used with either plaintext or encrypted file-sharing protocols, the approach is flexible and complete. The program has been trained and tested extensively with a wide variety of ransomware binaries and ‘not infected’ traffic, demonstrating that it is capable of recognizing all reported ransomware versions, including previously undiscovered ones. This research provides additional proof of the algorithm’s effectiveness by looking at the number of false positives and the amount of file encryption needed before discovery.

In another study ^[17], the author applies machine learning methods to the challenge of identifying maliciously encrypted messages. Detecting harmful encrypted traffic is a challenging problem, however, this paper ^[17] gives a thorough examination of the existing methods and datasets, as well as compares and contrasts numerous machine learning techniques. Several techniques and data sets are analyzed for their ability to detect such activity. The strengths and weaknesses of various machine learning methods for detecting malicious traffic in encrypted networks are compared and contrasted in this study. Researchers and practitioners can utilize the data to better protect themselves from these dangers.

To restrict people from accessing their files or disclosing important information, ransomware is a particularly dangerous form of malware. The inability to access encrypted files is a major issue for ransomware victims. Binary analysis of malware can be helpful for learning about the encryption methods used by different varieties of ransomware. In this article ^[18], the author examined the ecosystem for detecting ransomware, including the criteria, factors, and tools utilized in the process, and made comparisons between various methods and techniques. Researchers have also proposed a ransomware indexing system to better facilitate search, similarity checking, sample categorization, and clustering. The technique identifies native ransomware binaries using hybrid data from a static analyzer. By providing businesses with useful data and advice, this strategy enhances their ransomware defense.

A study ^[19] proposed a cloud-based method for classifying zero-day attacks using ML algorithms and cloud services. This research made use of Amazon Web Services to train and evaluate ML algorithms using a novel anomaly detection dataset, UGRansome1819. Three Machine Learning (ML) algorithms—Naive Bayes, Random Forest, and Support Vector Machine—are used in the proposed method of Ensemble Learning with a Genetic Algorithm optimizer. The terms “accuracy”, “F1-score”, “confusion matrix”, “recall”, and “precision” were all used to describe different aspects of a system’s performance. The results of the experiments demonstrate that UGRansome1819 provides superior classification accuracy to previously utilized datasets. The Genetic Algorithm can be used to pick features in order to reduce computational effort and prevent inappropriate model fitting. Classification accuracy can be enhanced by using an ensemble of classifiers, as in the optimum validation approach. The optimization procedure improves the accuracy of the SVM model and leads to high levels of specificity and sensitivity.

Cloud security is a vital area of study because of the challenges posed by cloud resource sharing, outsourcing, and multi-tenancy. New security issues have emerged due to the widespread adoption of web-based and trusted third-party technologies employed in the provision of cloud services. Despite significant progress accomplished in developing security models, procedures, and regulations, there are still obstacles to detecting new or unknown attacks and improving detection accuracy in the cloud. Deep learning (DL) and transfer learning (TL) are two examples of machine learning approaches that have been used to improve cloud security within these constraints ^[9]^[20]. Automatic and precise classification of safe and risky data is now achievable with the help of machine learning techniques. As a subset of machine learning, deep learning (DL) has proven particularly effective at addressing cloud security issues. In-depth discussions on cloud security, from the most basic precautions to the most advanced AI-based defenses. In addition to finding security holes in the cloud, it also assesses machine learning and DL-based security solutions and offers cutting-edge ways of managing vulnerabilities and threats.

Locking data with cryptographic methods, the ransomware then demands payment to unlock it. Current security measures are tested in the face of zero-day ransomware attacks, which exploit previously unknown flaws. Zero-shot Learning (ZSL) allows us to deal with classes we have no prior experience in a safe manner when there is no time to collect training data before an attack. ZSL uses a combination of deep learning (DL) and transfer learning (TL) to achieve this. In this study, the researchers introduced the Deep Contractive Autoencoder-based Attribute Learning (DCAE-ZSL) method and its counterpart, the Heterogeneous Voting Ensemble (HVE) Inference Stage (IS) ^[21]. DCAE-ZSL employs a Contractive Autoencoder (CAE) to extract fundamental characteristics of both known and unknown malware, while the IS aggregates many voting criteria to arrive at a final prediction. Models that are trained with contractive embeddings perform well against zero-day attacks, according to empirical evidence. The suggested voting-based ensemble (DCAE-ZSL-HVE) uses these essential features to better detection of zero-day attacks (recall = 0.95, FN = 6).

Infection with ransomware, a type of malicious software, can have devastating effects on its victims. The severity of these assaults can be reduced if they are discovered as soon as feasible. None of the numerous studies that have looked at the history, classification, current risks, and potential countermeasures of ransomware have addressed the need for dynamic analysis for ransomware detection across platforms. In order to address this gap in knowledge, the research work ^[1] looked into the datasets employed by cross-platform ransomware detection studies. As an added bonus, it provides a concise overview of research on ransomware detection strategies that make use of dynamic analysis, machine learning, deep learning, and hybrid approaches. By examining the subject of ransomware detection from the viewpoint of dynamic analytic methodologies, this study offers a fresh viewpoint on the topic.

Encrypted communications have become the standard on the internet as a result of the increased concern for personal privacy and sensitive information. While there are numerous good uses for encryption, it is also being used by bad actors to hide their own misdeeds from the public eye. Especially after COVID-19, when malicious encrypted traffic became increasingly common, this is of paramount importance. Deep packet inspection and other traditional security approaches are rendered ineffective due to the impossibility of doing simple payload content analysis. In order to detect bogus encrypted messages, it is crucial to employ machine learning-based methods. In this study ^[22], the authors conducted a systematic review of the methods currently in use to use machine learning to detect harmful encrypted communication. Recent research has struggled to provide fair comparisons of model performance due to the use of varying datasets. Because of this, the author merged data from five sources to create a more comprehensive and accurate dataset for use in future studies. The suggested system employs transfer learning (TL) and deep learning (DL) methods in an effort to improve the detection of ransomware assaults on cloud-encrypted data.

Using machine learning techniques, the following articles ^[8]^[23] demonstrate their ability to accurately detect ransomware in a local cloud setting (Table 1). In order to boost the reliability of ransomware detection, the study suggests using meta-features retrieved from volatile memory. Taking these meta-features into consideration, the proposed method aims to provide a reliable and quick method of detecting ransomware attacks in a private cloud environment. The fundamental contribution of this research is a methodology for extracting meta-features from RAM that can be used to enhance ransomware detection. In an effort to improve ransomware defense strategies in private cloud environments, the findings may be valuable for researchers and practitioners alike.

Table 1. Comparison of previous studies.

Reference	Techniques	Limitations	Key Contributions
^[1]	Dynamic analysis, machine learning	Limited exploration of research directions	Survey and research directions on ransomware detection
^[2]	Transfer learning, optimized CNN	Limited to IoT devices, no comparative evaluation	Transfer learning approach to IDS on Cloud IoT devices
^[3]	Deep learning	Limited to the IoT environment, no comparative evaluation	Optimal deep learning-based ransomware detection and classification in IoT environment
^[6]	Automated detection techniques	Limited to the IoT environment, no comparative evaluation	Trends and future directions in automated ransomware detection
^[10]	Review and analysis of existing methods	Limited to review and future directions, no new technique proposed	Review and future directions for ransomware detection, avoidance, and mitigation
^[24]	File entropy analysis, machine learning	Limited to file entropy analysis, no comparison with other techniques	Enhancing file entropy analysis to improve the machine learning detection rate of ransomware
^[16]	Machine learning models, encrypted traffic analysis	Limited to file-sharing network scenarios	Detection of crypto-ransomware using machine learning models in file-sharing network scenarios with encrypted traffic
^[12]	Machine learning, volatile memory analysis	Limited to a private cloud environment	Trusted detection of ransomware in a private cloud using machine learning methods leveraging meta-features from volatile memory
^[20]	Deep contractive autoencoder, the ensemble classifier	Limited to zero-day ransomware attacks	Zero-day ransomware attack detection using deep contractive autoencoder and voting-based ensemble classifier
^[21]	Machine learning, encrypted traffic analysis	Limited to comparative study, no specific limitations were mentioned	Comparative study of machine learning techniques for encrypted malicious traffic detection

The cybersecurity sector is in dire need of ransomware detection and prevention solutions ^[25]. Many tactics for protecting against ransomware attacks have been discussed in the academic literature ^[26]^[27]^[28]^[29]. Some of the methods that have been investigated by researchers for ransomware detection include dynamic analysis ^[30], machine learning algorithms ^[31], file entropy studies ^[32], deep learning models ^[33], and transfer learning ^[34]^[35]. Meta-features extracted from volatile memory and the analysis of encrypted traffic have both been investigated for their potential use in ransomware detection.

References

Urooj, U.; Al-Rimy, B.A.S.; Zainal, A.; Ghaleb, F.A.; Rassam, M.A. Ransomware Detection Using the Dynamic Analysis and Machine Learning: A Survey and Research Directions. Appl. Sci. 2022, 12, 172.
Okey, O.D.; Melgarejo, D.C.; Saadi, M.; Rosa, R.L.; Kleinschmidt, J.H.; Rodriguez, D.Z. Transfer Learning Approach to IDS on Cloud IoT Devices Using Optimized CNN. IEEE Access 2023, 11, 1023–1038.
Alohali, M.A.; Elsadig, M.; Al-Wesabi, F.N.; Al Duhayyim, M.; Hilal, A.M.; Motwakel, A. Optimal Deep Learning Based Ransomware Detection and Classification in the Internet of Things Environment. Comput. Syst. Sci. Eng. 2023, 46, 3087–3102.
Lee, K.; Lee, S.Y.; Yim, K. Machine Learning Based File Entropy Analysis for Ransomware Detection in Backup Systems. IEEE Access 2019, 7, 110205–110215.
Aslan, O.; Yilmaz, A.A. A New Malware Classification Framework Based on Deep Learning Algorithms. IEEE Access 2021, 9, 87936–87951.
Jegede, A.; Fadele, A.; Onoja, M.; Aimufua, G.; Mazadu, I.J. Trends and Future Directions in Automated Ransomware Detection. J. Comput. Soc. Inform. 2022, 1, 17–41.
Horduna, M.; Lazarescu, S.; Simion, E. A note on machine learning applied in ransomware detection. Int. Assoc. Cryptologic Res. 2023, 17. Available online: https://eprint.iacr.org/2023/045.pdf (accessed on 2 June 2023).
Bae, S.I.; Lee, G.B.; Im, E.G. Ransomware detection using machine learning algorithms. Concurr. Comput. Pract. Exp. 2020, 32, e5422.
Vehabovic, A.; Ghani, N.; Bou-Harb, E.; Crichigno, J.; Yayimli, A. Ransomware Detection and Classification Strategies. In Proceedings of the 2022 IEEE International Black Sea Conference on Communications and Networking (BlackSeaCom), Sofia, Bulgaria, 6–9 June 2022; pp. 316–324.
Apruzzese, G.; Laskov, P.; Montes de Oca, E.; Mallouli, W.; Brdalo Rapa, L.; Grammatopoulos, A.V.; Di Franco, F. The role of machine learning in cybersecurity. Digit. Threat. Res. Pract. 2023, 4, 1–38.
Gibert, D.; Planes, J.; Mateu, C.; Le, Q. Fusing feature engineering and deep learning: A case study for malware classification. Expert Syst. Appl. 2022, 207, 117957.
Berrueta, E.; Morato, D.; Magaña, E.; Izal, M. Crypto-ransomware detection using machine learning models in file-sharing network scenarios with encrypted traffic. Expert Syst. Appl. 2022, 209, 118299.
Kapoor, A.; Gupta, A.; Gupta, R.; Tanwar, S.; Sharma, G.; Davidson, I.E. Ransomware detection, avoidance, and mitigation scheme: A review and future directions. Sustainability 2021, 14, 8.
Al-Fawa’reh, M.; Ashi, Z.; Jafar, M.T. Detecting Malicious DNS Queries over Encrypted Tunnels Using Statistical Analysis and Bi-Directional Recurrent Neural Networks. Karbala Int. J. Mod. Sci. 2021, 7, 268–280.
Hsu, C.M.; Yang, C.C.; Cheng, H.H.; Setiasabda, P.E.; Leu, J.S. Enhancing File Entropy Analysis to Improve Machine Learning Detection Rate of Ransomware. IEEE Access 2021, 9, 138345–138351.
Smith, D.; Khorsandroo, S.; Roy, K. Machine Learning Algorithms and Frameworks in Ransomware Detection. IEEE Access 2022, 10, 117597–117610.
Cohen, A.; Nissim, N. Trusted detection of ransomware in a private cloud using machine learning methods leveraging meta-features from volatile memory. Expert Syst. Appl. 2018, 102, 158–178.
Yamany, B.; Elsayed, M.S.; Jurcut, A.D.; Abdelbaki, N.; Azer, M.A. A New Scheme for Ransomware Classification and Clustering Using Static Features. Electronics 2022, 11, 3307.
Nkongolo, M.; van Deventer, J.P.; Kasongo, S.M.; Zahra, S.R.; Kipongo, J. A Cloud Based Optimization Method for Zero-Day Threats Detection Using Genetic Algorithm and Ensemble Learning. Electronics 2022, 11, 1749.
Nenvani, G.; Gupta, H. A survey on attack detection on cloud using supervised learning techniques. In Proceedings of the 2016 Symposium on Colossal Data Analysis and Networking, CDAN 2016, Indore, India, 18–19 March 2016; Volume 175, pp. 21–27.
Zahoora, U.; Rajarajan, M.; Pan, Z.; Khan, A. Zero-day Ransomware Attack Detection using Deep Contractive Autoencoder and Voting based Ensemble Classifier. Appl. Intell. 2022, 52, 13941–13960.
Wang, Z.; Fok, K.W.; Thing, V.L.L. Machine learning for encrypted malicious traffic detection: Approaches, datasets and comparative study. Comput. Secur. 2022, 113, 102542.
Ren, A.L.Y.; Liang, C.T.; Hyug, I.J.; Brohi, S.N.; Jhanjhi, N.Z. A three-level ransomware detection and prevention mechanism. EAI Endorsed Trans. Energy Web 2020, 7, e6.
Fernando, D.W.; Komninos, N.; Chen, T. A Study on the Evolution of Ransomware Detection Using Machine Learning and Deep Learning Techniques. Internet Things 2020, 1, 551–604.
Ahanger, T.A.; Tariq, U.; Dahan, F.; Chaudhry, S.A.; Malik, Y. Securing IoT Devices Running PureOS from Ransomware Attacks: Leveraging Hybrid Machine Learning Techniques. Mathematics 2023, 11, 2481.
Sathya, T.; Keertika, N.; Shwetha, S.; Upodhyay, D.; Muzafar, H. Bitcoin Heist Ransomware Attack Prediction Using Data Science Process. E3S Web Conf. 2023, 399, 04056.
Alsaif, S.A. Machine Learning-Based Ransomware Classification of Bitcoin Transactions. Appl. Comput. Intell. Soft Comput. 2023, 2023, 6274260.
Sharma, T.; Patni, K.; Li, Z.; Trajković, L. Deep Echo State Networks for Detecting Internet Worm and Ransomware Attacks. In Proceedings of the 2023 IEEE International Symposium on Circuits and Systems (ISCAS), Monterey, CA, USA, 21–25 May 2023.
Thummapudi, K.; Lama, P.; Boppana, R.V. Detection of Ransomware Attacks using Processor and Disk Usage Data. IEEE Access 2023, 11, 51395–51407.
Ba’abbad, I.; Batarfi, O. Proactive Ransomware Detection Using Extremely Fast Decision Tree (EFDT) Algorithm: A Case Study. Computers 2023, 12, 121.
Charmilisri, A.; Harshi, I.; Madhushalini, V.; Raja, L. A Novel Ransomware Virus Detection Technique using Machine and Deep Learning Methods. In Proceedings of the 2023 7th International Conference on Intelligent Computing and Control Systems (ICICCS), Madurai, India, 17–19 May 2023.
Kumbhar, V.R.; Shende, A.P.; Raut, Y. Advance Model For Ransomware Attacking Data Classification And Prediction Using Ai. In Proceedings of the 2023 1st International Conference on Innovations in High Speed Communication and Signal Processing (IHCSP), Bhopal, India, 4–5 March 2023.
Almomani, I.; Alkhayer, A.; El-Shafai, W. E2E-RDS: Efficient End-to-End Ransomware Detection System Based on Static-Based ML and Vision-Based DL Approaches. Sensors 2023, 23, 4467.
Jin, B.; Cruz, L.; Gonçalves, N. Deep Facial Diagnosis: Deep Transfer Learning From Face Recognition to Facial Diagnosis. IEEE Access 2020, 8, 123649–123661.
Zhao, K.; Jia, F.; Shao, H. A novel conditional weighting transfer Wasserstein auto-encoder for rolling bearing fault diagnosis with multi-source domains. Knowl.-Based Syst. 2023, 262, 110203.

© Text is available under the terms and conditions of the Creative Commons Attribution (CC BY) license; additional terms may apply. By using this site, you agree to the Terms and Conditions and Privacy Policy.

Upload a video for this entry

Information

Subjects: Computer Science, Cybernetics; Computer Science, Information Systems

Contributors MDPI registered users' name will be linked to their SciProfiles pages. To register with us, please refer to https://encyclopedia.pub/register :

Amardeep Singh

Zohaib Mushtaq

Hamad Ali Abosaq

Salim Nasar Faraj Mursal

Muhammad Irfan

Grzegorz Nowakowski

View Times: 150

Update Date: 11 Dec 2023

Table of Contents

Video Upload Options

Confirm

1. Introduction

2. Enhancing Ransomware Attack Detection Using Transfer Learning and Deep Learning Ensemble Models on Cloud-Encrypted Data

References