With the advent of big data and cloud services, user data has become an important issue. Although a variety of detection and prevention technologies are used to protect user data, ransomware that demands money in exchange for one's data has emerged. In order to detect and prevent ransomware, file-and behavior-based detection methods have been investigated. Nevertheless, we are still facing from ransomware threats, as it is difficult to detect and prevent ransomware containing unknown malicious codes. In particular, these methods are limited in that they cannot detect ransomware for backup systems such as cloud services. For instance, if files infected with ransomware are synchronized with the backup systems, the infected files will not be able to be restored through the backed-up files. In this paper, we utilize an entropy technique to measure a characteristic of the encrypted file (i.e., uniformity). Machine learning is applied for classifying infected files based file entropy analysis. The proposed method can recover the original file from the backup system by detecting ransomware infected files that have been synchronized to the backup system, even if the user system is infected by ransomware. Conducted analysis results confirm that the proposed method provides a high detection rate with low false positive and false negative rates compared with the existing detection methods.

"Internet of Things" (IoT) bridges the communication barrier between the computing entities by forming a network between them. With a common solution for control and management of IoT devices, these networks are prone to all types of computing threats. Such networks may experience threats which are launched by exploitation of vulnerabilities that are left unhandled during the testing phases. These are often termed as "zero-day" vulnerabilities, and their conversion into a network attack is named as "zeroday" attack. These attacks can affect the IoT devices by exploiting the defense perimeter of the network. The existing solutions are capable of detecting such attacks but do not facilitate communication, which affects the performance of the network. In this paper, a consensus framework is proposed for mitigation of zero-day attacks in IoT networks. The proposed approach uses context behavior of IoT devices as a detection mechanism followed by alert message protocol and critical data sharing protocol for reliable communication during attack mitigation. The numerical analysis suggests that the proposed approach can serve the purpose of detection and elimination of zero-day attacks in IoT network without compromising its performance.

USB memory devices are both portable and easily accessible, and have thus become one of the most popular forms of external storage device. However, if a USB device is lost, stolen, or hacked, it may lead to leakage of critical information. It is a logical outcome that malicious individuals will try to steal their colleagues' USB memories. Consequently, various USB products with built-in security functions have been developed. To our knowledge, there has been little or no security analysis and comparison of these devices. This paper explores technological and architectural approaches to secure USB memories while analyzing their vulnerabilities, especially for resistance to reverse engineering attacks on the authentication protocols and data decryption. In this analysis, we classify vulnerabilities of these devices into 12 categories to summarize the current security situations on USB memories. Additionally, we derive a more secure authentication protocol based on our analysis. It is expected for secure USB products, including USB memory devices, to be revised with enhanced authentication protocols as a result of this effort.

Due to the emergence of online society, a representative user authentication method that is password authentication has been a key topic. However, in this authentication method, various attack techniques have emerged to steal passwords input from the keyboard, hence, the keyboard data does not ensure security. To detect and prevent such an attack, a keyboard data protection technique using random keyboard data generation has been presented. This technique protects keyboard data by generating dummy keyboard data while the attacker obtains the keyboard data. In this study, we demonstrate the feasibility of keyboard data exposure under the keyboard data protection technique. To prove the proposed attack technique, we gathered all the dummy keyboard data generated by the defense tool, and the real keyboard data input by the user, and evaluated the cybersecurity threat of keyboard data based on the machine learning-based offensive technique. We verified that an adversary obtains the keyboard data with 96.2% accuracy even if the attack technique that makes it impossible to attack keyboard data exposure is used. Namely, the proposed method in this study obviously differentiates the keyboard data input by the user from dummy keyboard data. Therefore, the contributions of this paper are that we derived and verified a new security threat and a new vulnerability of password authentication. Furthermore, a new cybersecurity threat derived from this study will have advantages over the security assessment of password authentication and all types of authentication technology and application services input from the keyboard.

An effective machine learning implementation means that artificial intelligence has tremendous potential to help and automate financial threat assessment for commercial firms and credit agencies. The scope of this study is to build a predictive framework to help the credit bureau by modelling/assessing the credit card delinquency risk. Machine learning enables risk assessment by predicting deception in large imbalanced data by classifying the transaction as normal or fraudster. In case of fraud transaction, an alert can be sent to the related financial organization that can suspend the release of payment for particular transaction. Of all the machine learning models such as RUSBoost, decision tree, logistic regression, multilayer perceptron, K-nearest neighbor, random forest, and support vector machine, the overall predictive performance of customized RUSBoost is the most impressive. The evaluation metrics used in the experimentation are sensitivity, specificity, precision, F scores, and area under receiver operating characteristic and precision recall curves. Datasets used for training and testing of the models have been taken from kaggle.com.