Biblio

Cyber-attacks on our Nation's Critical Infrastructure are growing. In this research, a Cyber Threat Intelligence (CTI) framework is proposed, developed, and tested. The results of the research, using 5 different simulated attacks on a dataset from an Industrial Control System (ICS) testbed, are presented with the extracted IOCs. The Bagging Decision Trees model showed the highest performance of testing accuracy (94.24%), precision (0.95), recall (0.93), and F1-score (0.94) among the 9 different machine learning models studied.

With the gradually popular high-speed wireless networks and 5G environments, the quality and reliability of network services will be suited for mobile vehicles. In addition to communicating information between vehicles, they can also communicate information with surrounding roadside equipment, pedestrians or traffic signs, and thus improve the road safety of passers-by.Recently, various countries have continuously invested in research on autonomous driving and unmanned vehicles. The open communication environment of the Internet of Vehicles in 5G will expose all personal information in the field of wireless networks. This research is based on the consideration of information security and personal data protection. We will focus on how to protect the real-time transmission of information between mobile vehicles to prevent from imbedding or altering important transmission information by unauthorized vehicles, drivers or passers-by participating in communications. Moreover, this research proposes a multi-level security key management agreement based on a dynamic M-tree structure for Internet of Vehicles to achieve flexible and scalable key management on large-scale Internet of Vehicles.

Security patterns are proven solutions to recurring problems in software development. The growing importance of secure software development has introduced diverse research efforts on security patterns that mostly focused on classification schemes, evolution and evaluation of the patterns. Despite a huge mature history of research and popularity among researchers, security patterns have not fully penetrated software development practices. Besides, software security education has not been benefited by these patterns though a commonly stated motivation is the dissemination of expert knowledge and experience. This is because the patterns lack a simple embodiment to help students learn about vulnerable code, and to guide new developers on secure coding. In order to address this problem, we propose to conduct intelligent data mining in the context of software engineering to discover learner-friendly software security patterns. Our proposed model entails knowledge discovery from large scale published real-world vulnerability histories in software applications. We harness association rule mining for frequent pattern discovery to mine easily comprehensible and explainable learner-friendly rules, mainly of the type "flaw implies fix" and "attack type implies flaw", so as to enhance training in secure coding which in turn would augment secure software development. We propose to build a learner-friendly intelligent tutoring system (ITS) based on the newly discovered security patterns and rules explored. We present our proposed model based on association rule mining in secure software development with the goal of building this ITS. Our proposed model and prototype experiments are discussed in this paper along with challenges and ongoing work.

Static information flow analysis (IFA) and dynamic information flow tracking (DIFT) have been widely employed in offline security analysis of computer programs. As security attacks become more sophisticated, there is a rising need for IFA and DIFT in production environment. However, existing systems usually deal with IFA and DIFT separately, and most DIFT systems incur significant performance overhead. We propose MIT to facilitate IFA and DIFT in online production environment. MIT offers mixed-mode information flow tracking at byte-granularity and incurs moderate runtime performance overhead. The core techniques consist of the extraction of taint semantics intermediate representation (TSIR) at compile-time and the decoupled execution of TSIR for information flow analysis. We conducted an extensive performance overhead evaluation on MIT to confirm its applicability in production environment. We also outline potential applications of MIT, including the implementation of data provenance checking and information flow based anomaly detection in real-world applications.

As cyber threats continue to grow and expertise resources are limited, organisations need to find ways to evaluate their resilience efficiently and take proactive measures against an attack from a specific adversary before it occurs. Threat modelling is an excellent method of assessing the resilience of ICT systems, forming Attack (Defense) Graphs (ADGs) that illustrate an adversary’s attack vectors. Cyber Threat Intelligence (CTI) is information that helps understand the current cyber threats, but has little integration with ADGs. This paper contributes with an approach that resolves this problem by using CTI feeds of known threat actors to enrich ADGs under multiple reuse. This enables security analysts to take proactive measures and strengthen their ICT systems against current methods used by any threat actor that is believed to pose a threat to them.

The extraordinary number of alerts generated by network intrusion detection systems (NIDS) can desensitize security analysts tasked with incident response. Security information and event management systems (SIEMs) perform some rudimentary automation but cannot replicate the decision-making process of a skilled analyst. Machine learning and artificial intelligence (AI) can detect patterns in data with appropriate training. In practice, the majority of the alert data comprises false alerts, and true alerts form only a small proportion. Consequently, a naive engine that classifies all security alerts into the majority class can yield a superficial high accuracy close to 100%. Without any correction for the class imbalance, the false alerts will dominate algorithmic predictions resulting in poor generalization performance. We propose a machine-learning approach to address the class imbalance problem in multi-appliance security alert data and automate the security alert analysis process performed in security operations centers (SOCs). We first used the neighborhood cleaning rule (NCR) to identify and remove ambiguous, noisy, and redundant false alerts. Then, we applied the support vector machine synthetic minority oversampling technique (SVMSMOTE) to generate synthetic training true alerts. Finally, we fit and evaluated the decision tree and random forest classifiers. In the experiments, using alert data from eight security appliances, we demonstrated that the proposed method can significantly reduce the need for manual auditing, decreasing the number of uninspected alerts and achieving a performance of 99.524% in recall.

Cybersecurity has become an emerging challenge for business information management and critical infrastructure protection in recent years. Artificial Intelligence (AI) has been widely used in different fields, but it is still relatively new in the area of Cyber-Physical Systems (CPS) security. In this paper, we provide an approach based on Machine Learning (ML) to intelligent threat recognition to enable run-time risk assessment for superior situation awareness in CPS security monitoring. With the aim of classifying malicious activity, several machine learning methods, such as k-nearest neighbours (kNN), Naïve Bayes (NB), Support Vector Machine (SVM), Decision Tree (DT) and Random Forest (RF), have been applied and compared using two different publicly available real-world testbeds. The results show that RF allowed for the best classification performance. When used in reference industrial applications, the approach allows security control room operators to get notified of threats only when classification confidence will be above a threshold, hence reducing the stress of security managers and effectively supporting their decisions.

In recent years, the development of deep learning has brought new ideas to internal threat detection. In this paper, three common deep learning algorithms for threat detection are optimized and innovated, and feature embedding, drift detection and sample weighting are introduced into FCNN. Adaptive multi-iteration method is introduced into Support Vector Data Description (SVDD). A dynamic threshold adjustment mechanism is introduced in VAE. In threat detection, three methods are used to detect the abnormal behavior of users, and the intersection of output results is taken as the final threat judgment basis. Experiments on cert r6.2 data set show that this method can significantly reduce the false positive rate.

Smart grid architecture and Software-defined Networking (SDN) have evolved into a centrally controlled infrastructure that captures and extracts data in real-time through sensors, smart-meters, and virtual machines. These advances pose a risk and increase the vulnerabilities of these infrastructures to sophisticated cyberattacks like distributed denial of service (DDoS), false data injection attack (FDIA), and Data replay. Integrating machine learning with a network intrusion detection system (NIDS) can improve the system's accuracy and precision when detecting suspicious signatures and network anomalies. Analyzing data in real-time using trained and tested hyperparameters on a network traffic dataset applies to most network infrastructures. The NSL-KDD dataset implemented holds various classes, attack types, protocol suites like TCP, HTTP, and POP, which are critical to packet transmission on a smart grid network. In this paper, we leveraged existing machine learning (ML) algorithms, Support vector machine (SVM), K-nearest neighbor (KNN), Random Forest (RF), Naïve Bayes (NB), and Bagging; to perform a detailed performance comparison of selected classifiers. We propose a multi-level hybrid model of SVM integrated with RF for improved accuracy and precision during network filtering. The hybrid model SVM-RF returned an average accuracy of 94% in 10-fold cross-validation and 92.75%in an 80-20% split during class classification.

Computer network and virtual machine security is very essential in today's era. Various architectures have been proposed for network security or prevent malicious access of internal or external users. Various existing systems have already developed to detect malicious activity on victim machines; sometimes any external user creates some malicious behavior and gets unauthorized access of victim machines to such a behavior system considered as malicious activities or Intruder. Numerous machine learning and soft computing techniques design to detect the activities in real-time network log audit data. KKDDCUP99 and NLSKDD most utilized data set to detect the Intruder on benchmark data set. In this paper, we proposed the identification of intruders using machine learning algorithms. Two different techniques have been proposed like a signature with detection and anomaly-based detection. In the experimental analysis, demonstrates SVM, Naïve Bayes and ANN algorithm with various data sets and demonstrate system performance on the real-time network environment.

Cloud computing is one of the greatest innovations and emerging technologies of the century. It incorporates networks, databases, operating systems, and virtualization technologies thereby bringing the security challenges associated with these technologies. Security Measures such as two-factor authentication, intrusion detection systems, and data backup are already in place to handle most of the security threats and vulnerabilities associated with these technologies but there are still other threats that may not be easily detected. Such a threat is a malicious user gaining access to the Virtual Machines (VMs) of other genuine users and using the Virtual Machine resources for their benefits without the knowledge of the user or the cloud service provider. This research proposes a model for proactive monitoring and detection of anomalies in VM resource usage. The proposed model can detect and pinpoint the time such anomaly occurred. Isolation Forest and One-Class Support Vector Machine (OCSVM) machine learning algorithms were used to train and test the model on sampled virtual machine workload trace using a combination of VM resource metrics together. OCSVM recorded an average F1-score of 0.97 and 0.89 for hourly and daily time series respectively while Isolation Forest has an average of 0.93 and 0.80 for hourly and daily time series. This result shows that both algorithms work for the model however OCSVM had a higher classification success rate than Isolation Forest.

It is not possible to speak about a complete security system without also taking into account the video surveillance system (CCTV). The reason is that CCTV systems offer the most spectacular results both in the security of goods and people and in the field of customer relations, marketing, traffic monitoring and the list can go on. With the development of the software industry the applicability of CCTV systems has greatly increased, largely due to image processing applications. The present paper, which is the continuation of an article already presented at an international conference, aims to shape the electrical characteristics of a common video surveillance system. The proposed method will be validated in two different programming environments.

Surveillance videos can capture anomalies in real scenarios and play an important role in security systems. Anomaly events are unpredictable, which reflect the unsupervised nature of the problem. In addition, it is difficult to construct a complete video dataset which contains all normal events. Based on the diversity of normal events, this paper proposes a memory-enhanced unsupervised method for anomaly detection. The proposed method reconstructs video events by combining prototype features and encoded features to detect anomaly events. Furthermore, a memory module is introduced to better store the prototype patterns of normal events. Experimental results in various benchmark datasets demonstrate the effectiveness and robustness of the proposed method.

Data security and privacy have become an important problem while big data systems are growing dramatically fast in various application fields. Paillier additive homomorphic cryptosystem is widely used in information security fields such as big data security, communication security, cloud computing security, and artificial intelligence security. However, how to improve its computational performance is one of the most critical problems in practice. In this paper, we propose two modifications to improve the performance of the Paillier cryptosystem. Firstly, we introduce a key generation method to generate the private key with low Hamming weight, and this can be used to accelerate the decryption computation of the Paillier cryptosystem. Secondly, we propose an acceleration method based on Hensel lifting in the Paillier cryptosystem. This method can obtain a faster and improved decryption process by showing the mathematical analysis of the decryption algorithm.

In recent days cloud domain gains a lot of user attention in order to store and access the data from remote locations connected through the internet. As it is generally known that all the sensitive data come from remote locations will be stored in the centralized storage medium and then try to access the data from that centralized storage space controlled by the cloud server. It is facing a problem like there is no security for the data in terms of user authorization and data authentication from the centralized storage location. Hence, it is required to migrate for a new storage procedure like Decentralized storage of cloud data in which the systems that do not rely on a central authority, so that the collusion resistance can be avoided by maintaining a global identifier. Here, the term de-centralized access means granting multi authorities to control the access for providing more security for the sensitive data. The proposed research study attempts to develop a new scheme by adding a global identifier like Attribute Authority (AA) for providing access keys for the data users who wish to access the sensitive information from the cloud server. The proposed research work attempts to incorporate the composite order bilinear groups scheme for providing access facility for the data users and provide more security for the sensitive data. By conducting various experiments on the proposed model, the obtained result clearly tells that the proposed system is very efficient to access the data in a de-centralized manner by using a global identifier.

To improve communication underwater, the underwater sensor networks (UWSN) provide gains for many different underwater applications, like Underwater Data-centers, Aquatic Monitoring, Tsunami Monitoring Systems, Aquatic Monitoring, Underwater Oil Field Discovery, Submarine Target Localization, Surveilling Water Territory of the Country via UWSN, Submarine Target Localization and many more. underwater applications are dependent on secure data communication in an underwater environment, so Data transmission in Underwater Sensor Network is a need of the future. Underwater data transmission itself is a big challenge due to various limitations of underwater communication mediums like lower bandwidth, multipath effect, path loss, propagation delay, noise, Doppler spread, and so on. These challenges make the underwater networks one of the most vulnerable networks for many different security attacks like sinkhole, spoofing, wormhole, misdirection, etc. It causes packets unable to be delivered to the destination, and even worse forward them to malicious nodes. A compromised node, which may be a router, intercepts packets going through it, and selectively drops them or can perform some malicious activity. This paper presents a solution to Mitigate unsecured data forwarding related attacks of an underwater sensor network, our solution uses a pre-shared key to secure communication and hashing algorithm to maintain the integrity of stored locations at head node and demonstration of attack and its mitigation done on Unetstack software.

Physical protection system (PPS) is developed to protect the assets or facilities against threats. A systematic analysis of the capabilities and intentions of potential threat capabilities is needed resulting in a so-called Design Basis Threat (DBT) document. A proper development of DBT is important to identify the system requirements that are required for adequately protecting a system and to optimize the resources needed for the PPS. In this paper we propose a model-based systems engineering approach for developing a DBT based on feature models. Based on a domain analysis process, we provide a metamodel that defines the key concepts needed for developing DBT. Subsequently, a reusable family feature model for PPS is provided that includes the common and variant properties of the PPS concepts detection, deterrence and response. The configuration processes are modeled to select and analyze the required features for implementing the threat scenarios. Finally, we discuss the integration of the DBT with the PPS design process.

Deepfake detection techniques have been widely studied to resolve security issues. However, existing techniques mainly focused on RGB channel-based analysis, which still shows incomplete detection accuracy. In this paper, we validate the performance of Gray channel-based deepfake detection. To compare RGB channel-based analysis and Gray channel-based analysis in deepfake detection, we quantitatively measured the performance by using popular CNN models, deepfake datasets, and evaluation indicators. Our experimental results confirm that Gray channel-based deepfake detection outperforms RGB channel-based deepfake detection in terms of accuracy and analysis time.

Video analytics has many applications in traffic control, security monitoring, action/event analysis, etc. With the adoption of deep neural networks, the accuracy of video analytics in video streams has been greatly improved. However, deep neural networks for performing video analytics are compute-intensive. In order to reduce processing time, many systems switch to the lower frame rate or resolution. State-of-the-art switching approaches adjust configurations by profiling video clips on a large configuration space. Multiple configurations are tested periodically and the cheapest one with a desired accuracy is adopted. In this paper, we propose a method that adapts the configuration by analyzing past video analytics results instead of profiling candidate configurations. Our method adopts a lower/higher resolution or frame rate when objects move slow/fast. We train a model that automatically selects the best configuration. We evaluate our method with two real-world video analytics applications: traffic tracking and pose estimation. Compared to the periodic profiling method, our method achieves 3%-12% higher accuracy with the same resource cost and 8-17x faster with comparable accuracy.

Nowadays, developing an intelligent system for segmenting the moving object from the background is essential task for video surveillance applications. Recently, a deep learning segmentation algorithm composed of encoder CNN, a Feature Pooling Module and a decoder CNN called FgSegNET\_S has been proposed. It is capable to train the model using few training examples. FgSegNET\_S is relying only on the spatial information while it is fundamental to include temporal information to distinguish if an object is moving or not. In this paper, an improved version known as (T\_FgSegNET\_S) is proposed by using the subtracted images from the initial background as input. The proposed approach is trained and evaluated using two publicly available infrared datasets: remote scene infrared videos captured by medium-wave infrared (MWIR) sensors and the Grayscale Thermal Foreground Detection (GTFD) dataset. The performance of network is evaluated using precision, recall, and F-measure metrics. The experiments show improved results, especially when compared to other state-of-the-art methods.

For exploiting multimode fiber optic communication networks towards physical layer security, we have trained a neural network performing mode decomposition of 10 modes. The approach is based on intensity-only camera images and works in real-time.

Recent cyber attacks on the power grid have been of increasing complexity and sophistication. In order to understand the impact of cyber-attacks on the power system resiliency, it is important to consider an holistic cyber-physical system specially with increasing industrial automation. In this study, device-level resilience properties of the various controllers and their impact on the microgrid resiliency is studied. In addition, a cyber-physical resiliency metric considering vulnerabilities, system model, and device-level properties is proposed. Resiliency is defined as the system ability to provide energy to critical loads even in extreme contingencies and depends on system ability to withstand, predict, and recover. A use case is presented inspired by the recent Ukraine cyber-attack. A use case has been presented to demonstrate application of the developed cyber-physical resiliency metric to enhance situational awareness of the operator, and enable better proactive or remedial control actions to improve resiliency.

With the advent of motorization, result in traffic system more congested, how to make the traffic system more effective and also take safety into account, namely build a intelligent transportation system, has become a hot spot of society. The vehicle distance control system studied in this paper is an important function in intelligent transportation system, through introducing cyber-physical systems (CPS) technology into it, set up system model, make the vehicles maintain a preset safety distance, thereby not only help improve the effective utilization of traffic system, but also help avoid the collision due to the speed change. Finally, use Simulink software to simulate and analyze the performance of the system, the result shows that the model can effectively cope with the distance change which is due to speed change, and ensure the vehicles maintain a preset safety distance within a short period of time.

Industry 4.0 connectivity requires ensuring end-to-end (E2E) security for industrial data. This requirement is critical when retrieving data from the OT network. Ciphertext-Policy Attribute-Based Encryption (CP-ABE) guarantees E2E security by encrypting data according to a policy and generating user keys according to attributes. To use this encryption scheme in manufacturing environments, policies must be updatable. This paper proposes a Multi-Layered Policy Key Encapsulation Method for CP-ABE that allows flexible policy update and revocation without modifying the original CP-ABE scheme.

Most famous malware defense tools depend on a large number of detect rules, which are time consuming to develop and require lots of professional experience. Meanwhile, even commercial tools may show high false-negative for some new coming malware, whose patterns were not curved in the prepared rules. This paper proposed the Normalized CNN based Malicious binary Detection method on condition of String, Feature mining (NCMDSF) to address the above problems. Firstly, amount of string feature was extracted from thousands of windows binary applications. Secondly, a 3-layer normalized CNN model, with normalization layer other than down sampling layer, was fit to detect malware. Finally, the proposed method NCMDSF was evaluated to discover malware from more than 1,000 windows binary applications by K-fold cross validation. Experimental results showed that, NCMDSF was superior to some other learning-based methods, including classical CNN, LSTM, normalized LSTM, and won higher true positive rate on the condition of same false positive rate. Furthermore, it successfully avoids over-fitting that occurs in deep learning methods without using normalization.