Biblio

We present a novel approach for the collaborative training of neural network models in decentralized federated environments. In the iterative process a group of autonomous peers run multiple training rounds to train a common model. Thereby, participants perform all model training steps locally, such as stochastic gradient descent optimization, using their private, e.g. mission-critical, training datasets. Based on locally updated models, participants can jointly determine a common model by averaging all associated model weights without sharing the actual weight values. For this purpose we introduce a simple n-out-of-n secret sharing schema and an algorithm to calculate average values in a peer-to-peer manner. Our experimental results with deep neural networks on well-known sample datasets prove the generic applicability of the approach, with regard to model quality parameters. Since there is no need to involve a central service provider in model training, the approach can help establish trustworthy collaboration platforms for businesses with high security and data protection requirements.

Code-reuse techniques have emerged as a way to defeat the control-flow defenses that prevent the injection and execution of new code, as they allow an adversary to hijack the control flow of a victim program without injected code. A well-known code-reuse attack technique is Return-OrientedProgramming (ROP), which considers and links together (relatively short) code snippets, named ROP gadgets, already present in the victim’s memory address space through a controlled use of the stack values of the victim program. Although ROP attacks are known to be Turing-complete, there are still open question such as the quantification of the executional power of an adversary, which is determined by whatever code exists in the memory of a victim program, and whether an adversary can build a ROP chain, made up of ROP gadgets, for any kind of algorithm. To fill these gaps, in this paper we first define a virtual language, dubbed ROPLANG, that defines a set of operations (specifically, arithmetic, assignment, dereference, logical, and branching operations) which are mapped to ROP gadgets. We then use it to evaluate the executional power of an adversary in Windows 7 and Windows 10, in both 32- and 64-bit versions. In addition, we have developed ROP3, a tool that accepts a set of program files and a ROP chain described with our language and returns the code snippets that make up the ROP chain. Our results show that there are enough ROP gadgets to simulate any virtual operation and that branching operations are the less frequent ones. As expected, our results also indicate that the larger a program file is, the more likely to find ROP gadgets within it for every virtual operation.

Safety- and security-critical developers have long recognized the importance of applying a high degree of scrutiny to a system’s (or subsystem’s) I/O messages. However, lack of care in the development of message-handling components can lead to an increase, rather than a decrease, in the attack surface. On the DARPA Cyber-Assured Systems Engineering (CASE) program, we have focused our research effort on identifying cyber vulnerabilities early in system development, in particular at the Architecture development phase, and then automatically synthesizing components that mitigate against the identified vulnerabilities from high-level specifications. This approach is highly compatible with the goals of the LangSec community. Advances in formal methods have allowed us to produce hardware/software implementations that are both performant and guaranteed correct. With these tools, we can synthesize high-assurance “building blocks” that can be composed automatically with high confidence to create trustworthy systems, using a method we call Security-Enhancing Architectural Transformations. Our synthesis-focused approach provides a higherleverage insertion point for formal methods than is possible with post facto analytic methods, as the formal methods tools directly contribute to the implementation of the system, without requiring developers to become formal methods experts. Our techniques encompass Systems, Hardware, and Software Development, as well as Hardware/Software Co-Design/CoAssurance. We illustrate our method and tools with an example that implements security-improving transformations on system architectures expressed using the Architecture Analysis and Design Language (AADL). We show how message-handling components can be synthesized from high-level regular or context-free language specifications, as well as a novel specification language for self-describing messages called Contiguity Types, and verified to meet arithmetic constraints extracted from the AADL model. Finally, we guarantee that the intent of the message processing logic is accurately reflected in the application binary code through the use of the verified CakeML compiler, in the case of software, or the Restricted Algorithmic C toolchain with ACL2-based formal verification, in the case of hardware/software co-design.

Ransomware is one of the most serious threats which constitute a significant challenge in the cybersecurity field. The cybercriminals use this attack to encrypts the victim's files or infect the victim's devices to demand ransom in exchange to restore access to these files and devices. The escalating threat of Ransomware to thousands of individuals and companies requires an urgent need for creating a system capable of proactively detecting and preventing ransomware. In this research, a new approach is proposed to detect and classify ransomware based on three machine learning algorithms (Random Forest, Support Vector Machines , and Näive Bayes). The features set was extracted directly from raw byte using static analysis technique of samples to improve the detection speed. To offer the best detection accuracy, CF-NCF (Class Frequency - Non-Class Frequency) has been utilized for generate features vectors. The proposed approach can differentiate between ransomware and goodware files with a detection accuracy of up to 98.33 percent.

Nowadays deepfake videos are concern with social ethics, privacy and security. Deepfake videos are synthetically generated videos that are generated by modifying the facial features and audio features to impose one person’s facial data and audio to other videos. These videos can be used for defaming and fraud. So, counter these types of manipulations and threats, detection of deepfake video is needed. This paper proposes multilayer hybrid recurrent deep learning models for deepfake video detection. Proposed models exploit the noise-based temporal facial convolutional features and temporal learning of hybrid recurrent deep learning models. Experiment results of these models demonstrate its performance over stacked recurrent deep learning models.

An electronic medical record (EMR) is the digital medical data of a patient, and they are healthcare system's most valuable asset. In this paper, we introduce a decentralized network using blockchain technology and smart contracts as a solution to manage and secure medical records storing, and transactions between medical healthcare providers. Ethereum blockchain is employed to build the blockchain. Solidity object-oriented language was utilized to implement smart contracts to digitally facilitate and verify transactions across the network (creating records, access requests, permitting access, revoking access, rejecting access). This will mitigate prevailing issues of current systems and enhance their performance, since current EMRs are stored on a centralized database, which cannot guarantee data integrity and security, consequently making them susceptible to malicious attacks. Our proposed system approach is of vital importance considering that healthcare providers depend on various tests in making a decision about a patient's diagnosis, and the respective plan of treatment they will go through. These tests are not shared with other providers, while data is scattered on various systems, as a consequence of these ensuing scenarios, patients suffer of the resulting care provided. Moreover, blockchain can meliorate the motley serious challenges caused by future use of IoT devices that provide real-time data from patients. Therefore, integrating the two technologies will produce decentralized IoT based healthcare systems.

Controller area network is the serial communication protocol, which broadcasts the message on the CAN bus. The transmitted message is read by all the nodes which shares the CAN bus. The message can be eavesdropped and can be re-used by some other node by changing the information or send it by duplicate times. The message reused after some delay is replay attack. In this paper, the CAN network with three CAN nodes is implemented using the universal verification components and the replay attack is demonstrated by creating the faulty node. Two types of replay attack are implemented in this paper, one is to replay the entire message and the other one is to replay only the part of the frame. The faulty node uses the first replay attack method where it behaves like the other node in the network by duplicating the identifier. CAN frame except the identifier is reused in the second method which is hard to detect the attack as the faulty node uses its own identifier and duplicates only the data in the CAN frame.

In the crowdsourced testing system, due to the openness of crowdsourced testing platform and other factors, the security of crowdsourced testing intellectual property cannot be effectively protected. We proposed an attribute-based double encryption scheme, combined with the blockchain technology, to achieve the data access control method of the code to be tested. It can meet the privacy protection and traceability of specific intellectual property in the crowdsourced testing environment. Through the experimental verification, the access control method is feasible, and the performance test is good, which can meet the normal business requirements.

Intrusion Detection System (IDS) is one of the most important security tool for many security issues that are prevailing in today's cyber world. Intrusion Detection System is designed to scan the system applications and network traffic to detect suspicious activities and issue an alert if it is discovered. So many techniques are available in machine learning for intrusion detection. The main objective of this project is to apply machine learning algorithms to the data set and to compare and evaluate their performances. The proposed application has used the SVM (Support Vector Machine) and ANN (Artificial Neural Networks) Algorithms to detect the intrusion rates. Each algorithm is used to detect whether the requested data is authorized or contains any anomalies. While IDS scans the requested data if it finds any malicious information it drops that request. These algorithms have used Correlation-Based and Chi-Squared Based feature selection algorithms to reduce the dataset by eliminating the useless data. The preprocessed dataset is trained and tested with the models to obtain the prominent results, which leads to increasing the prediction accuracy. The NSL KDD dataset has been used for the experimentation. Finally, an accuracy of about 48% has been achieved by the SVM algorithm and 97% has been achieved by ANN algorithm. Henceforth, ANN model is working better than the SVM on this dataset.

Commodity I/O hardware often fails to separate I/O transfers of isolated OS and applications code. Even when using the best I/O hardware, commodity systems sometimes trade off separation assurance for increased performance. Remarkably, device firmware need not be malicious. Instead, any malicious driver, even if isolated in its own execution domain, can manipulate its device to breach I/O separation. To prevent such vulnerabilities with high assurance, a formal I/O separation model and its use in automatic generation of secure I/O kernel code is necessary.This paper presents a formal I/O separation model, which defines a separation policy based on authorization of I/O transfers and is hardware agnostic. The model, its refinement, and instantiation in the Wimpy kernel design, are formally specified and verified in Dafny. We then specify the kernel implementation and automatically generate verified-correct assembly code that enforces the I/O separation policies. Our formal modeling enables the discovery of heretofore unknown design and implementation vulnerabilities of the original Wimpy kernel. Finally, we outline how the model can be applied to other I/O kernels and conclude with the key lessons learned.

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.

Protecting privacy of the user location in Internet of Things (IoT) applications is a complex problem. Peer-to-peer (P2P) approach is one of the most popular techniques used to protect privacy in IoT applications, especially that use the location service. The P2P approach requires trust among peers in addition to serious cooperation. These requirements are still an open problem for this approach and its methods. In this paper, we propose an effective solution to this issue by creating a manager for the peers' reputation called R-TTP. Each peer has a new query. He has to evaluate the cooperated peer. Depending on the received result of that evaluation, the main peer will send multiple copies of the same query to multiple peers and then compare results. Moreover, we proposed another scenario to the manager of reputation by depending on Fog computing to enhance both performance and privacy. Relying on this work, a user can determine the most suitable of many available cooperating peers, while avoiding the problems of putting up with an inappropriate cooperating or uncommitted peer. The proposed method would significantly contribute to developing most of the privacy techniques in the location-based services. We implemented the main functions of the proposed method to confirm its effectiveness, applicability, and ease of application.

Malware poses a threat to computing systems worldwide, and security experts work tirelessly to detect and classify malware as accurately and quickly as possible. Since malware can use evasion techniques to bypass static analysis and security mechanisms, dynamic analysis methods are more useful for accurately analyzing the behavioral patterns of malware. Previous studies showed that malware behavior can be represented by sequences of executed system calls and that machine learning algorithms can leverage such sequences for the task of malware classification (a.k.a. malware categorization). Accurate malware classification is helpful for malware signature generation and is thus beneficial to antivirus vendors; this capability is also valuable to organizational security experts, enabling them to mitigate malware attacks and respond to security incidents. In this paper, we propose an improved methodology for malware classification, based on analyzing sequences of system calls invoked by malware in a dynamic analysis environment. We show that adding an attention mechanism to a LSTM model improves accuracy for the task of malware classification, thus outperforming the state-of-the-art algorithm by up to 6%. We also show that the transformer architecture can be used to analyze very long sequences with significantly lower time complexity for training and prediction. Our proposed method can serve as the basis for a decision support system for security experts, for the task of malware categorization.

The Internet-of-Things (IoT) paradigm at large continues to be compromised, hindering the privacy, dependability, security, and safety of our nations. While the operational security communities (i.e., CERTS, SOCs, CSIRT, etc.) continue to develop capabilities for monitoring cyberspace, tools which are IoT-centric remain at its infancy. To this end, we address this gap by innovating an actionable Cyber Threat Intelligence (CTI) feed related to Internet-scale infected IoT devices. The feed analyzes, in near real-time, 3.6TB of daily streaming passive measurements ( ≈ 1M pps) by applying a custom-developed learning methodology to distinguish between compromised IoT devices and non-IoT nodes, in addition to labeling the type and vendor. The feed is augmented with third party information to provide contextual information. We report on the operation, analysis, and shortcomings of the feed executed during an initial deployment period. We make the CTI feed available for ingestion through a public, authenticated API and a front-end platform.

Mobile payment protocols have attracted widespread attention over the past decade, due to advancements in digital technology. The use of these protocols in online industries can dramatically improve the quality of online services. However, the central issue of concern when utilizing these types of systems is their accountability, which ensures trust between the parties involved in payment transactions. It is, therefore, vital for researchers to investigate how to handle the accountability of mobile payment protocols. In this research, we introduce a secure mobile payment protocol to overcome this problem. Our payment protocol combines all the necessary security features, such as confidentiality, integrity, authentication, and authorization that are required to build trust among parties. In other words, is the properties of mutual authentication and non-repudiation are ensured, thus providing accountability. Our approach can resolve any conflicts that may arise in payment transactions between parties. To prove that the proposed protocol is correct and complete, we use the Scyther and AVISPA tools to verify our approach formally.

IoT networks today face a myriad of security vulnerabilities in their infrastructure due to its wide attack surface. Large-scale networks are increasingly adopting a Software-Defined Networking approach, it allows for simplified network control and management through network virtualization. Since traditional security mechanisms are incapable of handling virtualized environments, SDSec or Software-Defined Security is introduced as a solution to support virtualized infrastructure, specifically aimed at providing security solutions to SDN frameworks. To further aid large scale design and development of SDN frameworks, Model-Driven Engineering (MDE) has been proposed to be used at the design phase, since abstraction, automation and analysis are inherently key aspects of MDE. This provides an efficient approach to reducing large problems through models that abstract away the complex technicality of the total system. Making adaptations to these models to address security issues faced in IoT networks, largely reduces cost and improves efficiency. These models can be simulated, analysed and supports architecture model adaptation; model changes are then reflected back to the real system. We propose a model-driven security approach for SDSec networks that can self-adapt using machine learning to mitigate security threats. The overall design time changes can be monitored at run time through machine learning techniques (e.g. deep, reinforcement learning) for real time analysis. This approach can be tested in IoT simulation environments, for instance using the CAPS IoT modeling and simulation framework. Using self-adaptation of models and advanced machine learning for data analysis would ensure that the SDSec architecture adapts and improves over time. This largely reduces the overall attack surface to achieve improved end-to-end security in IoT environments.

In the past few years, the blockchain emerged as peer-to-peer distributed ledger technology for recording transactions, maintained by many peers without any central trusted regulatory authority through distributed public-key cryptography and consensus mechanism. It has not only given the birth of cryptocurrencies, but it also resolved various security, privacy and transparency issues of decentralized systems. This article discussed the blockchain basics overview, architecture, and blockchain security components such as hash function, Merkle tree, digital signature, and Elliptic curve cryptography (ECC). In addition to the core idea of blockchain, we focus on ECC's significance in the blockchain. We also discussed why RSA and other key generation mechanisms are not suitable for blockchain-based IoT applications. We also analyze many possible blockchain-based applications where ECC algorithm is better than other algorithms concerning security and privacy assurance. At the end of the article, we will explain the comparative analysis of ECC and RSA.

This paper addresses security and risk management of hardware and embedded systems across several applications. There are three companies involved in the research. First is an energy technology company that aims to leverage electric- vehicle batteries through vehicle to grid (V2G) services in order to provide energy storage for electric grids. Second is a defense contracting company that provides acquisition support for the DOD's conventional prompt global strike program (CPGS). These systems need protections in their production and supply chains, as well as throughout their system life cycles. Third is a company that deals with trust and security in advanced logistics systems generally. The rise of interconnected devices has led to growth in systems security issues such as privacy, authentication, and secure storage of data. A risk analysis via scenario-based preferences is aided by a literature review and industry experts. The analysis is divided into various sections of Criteria, Initiatives, C-I Assessment, Emergent Conditions (EC), Criteria-Scenario (C-S) relevance and EC Grouping. System success criteria, research initiatives, and risks to the system are compiled. In the C-I Assessment, a rating is assigned to signify the degree to which criteria are addressed by initiatives, including research and development, government programs, industry resources, security countermeasures, education and training, etc. To understand risks of emergent conditions, a list of Potential Scenarios is developed across innovations, environments, missions, populations and workforce behaviors, obsolescence, adversaries, etc. The C-S Relevance rates how the scenarios affect the relevance of the success criteria, including cost, schedule, security, return on investment, and cascading effects. The Emergent Condition Grouping (ECG) collates the emergent conditions with the scenarios. The generated results focus on ranking Initiatives based on their ability to negate the effects of Emergent Conditions, as well as producing a disruption score to compare a Potential Scenario's impacts to the ranking of Initiatives. The results presented in this paper are applicable to the testing and evaluation of security and risk for a variety of embedded smart devices and should be of interest to developers, owners, and operators of critical infrastructure systems.

The ongoing trend of moving data and computation to the cloud is met with concerns regarding privacy and protection of intellectual property. Cloud Service Providers (CSP) must be fully trusted to not tamper with or disclose processed data, hampering adoption of cloud services for many sensitive or critical applications. As a result, CSPs and CPU manufacturers are rushing to find solutions for secure and trustworthy outsourced computation in the Cloud. While enclaves, like Intel SGX, are strongly limited in terms of throughput and size, AMD’s Secure Encrypted Virtualization (SEV) offers hardware support for transparently protecting code and data of entire VMs, thus removing the performance, memory and software adaption barriers of enclaves. Through attestation of boot code integrity and means for securely transferring secrets into an encrypted VM, CSPs are effectively removed from the list of trusted entities. There have been several attacks on the security of SEV, by abusing I/O channels to encrypt and decrypt data, or by moving encrypted code blocks at runtime. Yet, none of these attacks have targeted the attestation protocol, the core of the secure computing environment created by SEV. We show that the current attestation mechanism of Zen 1 and Zen 2 architectures has a significant flaw, allowing us to manipulate the loaded code without affecting the attestation outcome. An attacker may abuse this weakness to inject arbitrary code at startup–and thus take control over the entire VM execution, without any indication to the VM’s owner. Our attack primitives allow the attacker to do extensive modifications to the bootloader and the operating system, like injecting spy code or extracting secret data. We present a full end-to-end attack, from the initial exploit to leaking the key of the encrypted disk image during boot, giving the attacker unthrottled access to all of the VM’s persistent data.

Noise has been used as a way of protecting privacy of users in public datasets for many decades now. Differential privacy is a new standard to add noise, so that user privacy is protected. When this technique is applied for a single end user data, it's called local differential privacy. In this study, we evaluate the effects of adding noise to generate randomized responses on machine learning models. We generate randomized responses using Gaussian, Laplacian noise on singular end user data as well as correlated end user data. Finally, we provide results that we have observed on a few data sets for various machine learning use cases.

Recently, federated learning (FL), as an advanced and practical solution, has been applied to deal with privacy-preserving issues in distributed multi-party federated modeling. However, most existing FL methods focus on the same privacy-preserving budget while ignoring various privacy requirements of participants. In this paper, we for the first time propose an algorithm (PLU-FedOA) to optimize the deep neural network of horizontal FL with personalized local differential privacy. For such considerations, we design two approaches: PLU, which allows clients to upload local updates under differential privacy-preserving of personally selected privacy level, and FedOA, which helps the server aggregates local parameters with optimized weight in mixed privacy-preserving scenarios. Moreover, we theoretically analyze the effect on privacy and optimization of our approaches. Finally, we verify PLU-FedOA on real-world datasets.

Recent years have seen an increase in medical big data, which can be attributed to a paradigm shift experienced in medical data sharing induced by the growth of medical technology and the Internet of Things. The evidence of this potential has been proved during the recent covid-19 pandemic, which was characterised by the use of medical wearable devices to help with the medical data exchange between the healthcare providers and patients in a bid to contain the pandemic. However, the use of these technologies has also raised questions and concerns about security and privacy risks. To assist in resolving this issue, this paper proposes a blockchain-based access control framework for managing access to users’ medical data. This is facilitated by using a smart contract on the blockchain, which allows for delegated access control and secure user authentication. This solution leverages blockchain technology’s inherent autonomy and immutability to solve the existing access control challenges. We have presented the solution in the form of a medical wearable sensor prototype and a mobile app that uses the Ethereum blockchain in a real data sharing control scenario. Based on the empirical results, the proposed solution has proven effective. It has the potential to facilitate reliable data exchange while also protecting sensitive health information against potential threats. When subjected to security analysis and evaluation, the system exhibits performance improvements in data privacy levels, high security and lightweight access control design compared to the current centralised access control models.

Wireless Sensor Networks (WSNs) and their implementations have been the subject of numerous studies over the last two decades. WSN gathers, processes, and distributes wireless data to the database storage center. This study aims to explain the four main components of sensor nodes and the mechanism of WSN's. WSNs have 5 available types that will be discussed and explained in this paper. In addition to that, shortest path routing will be thoroughly analyzed. In “The Protocol”. Reconfigurable logic applications have grown in number and complexity. Shortest path routing is a method of finding paths through a network with the least distance or other cost metric. The efficiency of the shortest path protocol mechanism and the reliability of encryption are both present which adds security and accuracy of location privacy and message delivery. There are different forms of key management, such as symmetric and asymmetric encryption, each with its own set of processing techniques. The use of encryption technique to secure sensor nodes is addressed, as well as how we overcame the problem with the aid of advanced techniques. Our major findings are that adding more security doesn't cost much and by cost we mean energy consumption, throughput and latency.

This article describes a research work in progress, in which a methodology for the development of computerized educational materials based on augmented reality is proposed. The development of the proposal is preceded by a systematic review of the literature in which the convenience of having a methodology that assists teachers and developers interested in the development of educational materials related to augmented reality technology is concluded. The proposed methodology consists of four stages: (1) initiation, (2) design of the learning scenario, (3) implementation and (4) evaluation, as well as specific elements that must be considered in each of them for their correct fulfillment. Finally, the article briefly describes the validation strategy designed to evaluate this methodological proposal.

Multimedia Internet of Things (IoT) devices, especially, the smartphones are embedded with sensors including Global Positioning System (GPS), barometer, microphone, accelerometer, etc. These sensors working together, present a fairly complete picture of the citizens' daily activities, with implications for their privacy. With the internet, Citizens in Smart Cities are able to perform their daily life activities online with their connected electronic devices. But, unfortunately, computer hackers tend to write automated malicious applications to attack websites on which these citizens perform their activities. These security threats sometime put their private information at risk. In order to prevent these security threats on websites, Completely Automated Public Turing test to tell Computers and Humans Apart (CAPTCHAs) are generated, as a form of security mechanism to protect the citizens' private information. But with the advancement of deep learning, text-based CAPTCHAs can sometimes be vulnerable. As a result, it is essential to conduct performance evaluation on the CAPTCHAs that are generated before they are deployed on multimedia web applications. Therefore, this work proposed an optimal and light-weight Convolutional Neural Network (CNN) to solve both numerical and alpha-numerical complex text-based CAPTCHAs simultaneously. The accuracy of the proposed CNN model has been accelerated based on Cyclical Learning Rates (CLRs) policy. The proposed CLR-CNN model achieved a high accuracy to solve both numerical and alpha-numerical text-based CAPTCHAs of 99.87% and 99.66%, respectively. In real-time, we observed that the speed of the model has increased, the model is lightweight, stable, and flexible as compared to other CAPTCHA solving techniques. The result of this current work will increase awareness and will assist multimedia security Researchers to continue and develop more robust text-based CAPTCHAs with their security mechanisms capable of protecting the private information of citizens in Smart Cities.