External Attacks Research Articles

Security for Machine Learning-based Software Systems: A Survey of Threats, Practices, and Challenges

The rapid development of Machine Learning (ML) has demonstrated superior performance in many areas, such as computer vision and video and speech recognition. It has now been increasingly leveraged in software systems to automate the core tasks. However, how to securely develop the machine learning-based modern software systems (MLBSS) remains a big challenge, for which the insufficient consideration will largely limit its application in safety-critical domains. One concern is that the present MLBSS development tends to be rushed, and the latent vulnerabilities and privacy issues exposed to external users and attackers will be largely neglected and hard to be identified. Additionally, machine learning-based software systems exhibit different liabilities towards novel vulnerabilities at different development stages from requirement analysis to system maintenance, due to its inherent limitations from the model and data and the external adversary capabilities. The successful generation of such intelligent systems will thus solicit dedicated efforts jointly from different research areas, i.e., software engineering, system security, and machine learning. Most of the recent works regarding the security issues for ML have a strong focus on the data and models, which has brought adversarial attacks into consideration. In this work, we consider that security for machine learning-based software systems may arise from inherent system defects or external adversarial attacks, and the secure development practices should be taken throughout the whole lifecycle. While machine learning has become a new threat domain for existing software engineering practices, there is no such review work covering the topic. Overall, we present a holistic review regarding the security for MLBSS, which covers a systematic understanding from a structure review of three distinct aspects in terms of security threats. Moreover, it provides a thorough state-of-the-practice for MLBSS secure development. Finally, we summarize the literature for system security assurance and motivate the future research directions with open challenges. We anticipate this work provides sufficient discussion and novel insights to incorporate system security engineering for future exploration.

The acid and chloride permeability resistance of masonry cement plaster mortar incorporating high-volume fly ash content

The utilization of masonry cement mortar has been on the rise in recent years, gradually supplanting the traditional use of lime and Portland cement in plastering applications. Masonry cement mortar is favored for its enhanced workability; however, it remains vulnerable to environmental factors, and there is a dearth of information regarding the durability assessment of masonry cement plaster mortar. While incorporating fly ash in mortars is a common practice, it is often paired with Ordinary Portland Cement as the primary binding agent. This research project was devised to investigate the performance characteristics of masonry cement plaster mortar infused with varying levels of fly ash. The control specimen was systematically replaced with fly ash in 10% weight increments, up to 50%. Evaluations were conducted immediately in its fresh state, and assessments of strength and durability were extended over 360 days. Scanning Electron Microscope (SEM) and X-Ray diffraction analysis (XRD) were used as characterization techniques. The outcomes of the study were noteworthy. Mortars containing 30% or more ash exhibited a significant 15% enhancement in flow and an impressive 88% increase in compressive strength compared to plain mortar. The durability assessment revealed that masonry cement mortar incorporating 40% and 50% fly ash displayed nearly 14 times greater resistance to chloride penetration and maintained its strength when exposed to acidic solutions. This underscores the dual benefits of this material, enhancing the aesthetics of buildings and fortifying them against external chemical attacks that can compromise the integrity of concrete and bricks.

Implementing quantum anonymous multi-party ranking: the influence and application of the law of large numbers

Secure multi-party ranking is not only a pivotal component within the domain of secure multi-party computation but also holds extensive real-world applicability, given its efficacy in safeguarding the integrity and privacy of data. In this paper, we propose a verifiable quantum anonymous multi-party ranking protocol, which leverages the power of multi-particle entangled states in conjunction with the law of large numbers. The protocol achieves the secure ranking of participants’ private data while verifying their identities through the assistance of a semi-honest third party. We prove that the protocol is resistant to different types of attacks from internal or external attackers. In addition, the proposed protocol employs qubits as the information carriers, which improves the feasibility of the protocol. We demonstrate the feasibility of the protocol by using the online analog quantum computer of IBM Corporation placed on the cloud. Our research represents an innovative initiative that merges the realms of quantum cryptography and statistical analysis to address pertinent practical challenges.

Leveraging Artificial Intelligence and Provenance Blockchain Framework to Mitigate Risks in Cloud Manufacturing in Industry 4.0

Cloud manufacturing is an evolving networked framework that enables multiple manufacturers to collaborate in providing a range of services, including design, development, production, and post-sales support. The framework operates on an integrated platform encompassing a range of Industry 4.0 technologies, such as Industrial Internet of Things (IIoT) devices, cloud computing, Internet communication, big data analytics, artificial intelligence, and blockchains. The connectivity of industrial equipment and robots to the Internet opens cloud manufacturing to the massive attack risk of cybersecurity and cyber crime threats caused by external and internal attackers. The impacts can be severe because the physical infrastructure of industries is at stake. One potential method to deter such attacks involves utilizing blockchain and artificial intelligence to track the provenance of IIoT devices. This research explores a practical approach to achieve this by gathering provenance data associated with operational constraints defined in smart contracts and identifying deviations from these constraints through predictive auditing using artificial intelligence. A software architecture comprising IIoT communications to machine learning for comparing the latest data with predictive auditing outcomes and logging appropriate risks was designed, developed, and tested. The state changes in the smart ledger of smart contracts were linked with the risks so that the blockchain peers can detect high deviations and take actions in a timely manner. The research defined the constraints related to physical boundaries and weightlifting limits allocated to three forklifts and showcased the mechanisms of detecting risks of breaking these constraints with the help of artificial intelligence. It also demonstrated state change rejections by blockchains at medium and high-risk levels. This study followed software development in Java 8 using JDK 8, CORDA blockchain framework, and Weka package for random forest machine learning. As a result of this, the model, along with its design and implementation, has the potential to enhance efficiency and productivity, foster greater trust and transparency in the manufacturing process, boost risk management, strengthen cybersecurity, and advance sustainability efforts.

PREEN: An Aggregation Mechanism for Privacy-Preserving Smart-Metering Communications

The widespread use of home consumer eletronics (CEs) provides residents facility, assistance and comfort. Meanwhile, smart meters are deployed at households to optimize electricity supply and demand and improve electricity efficiency. These meters monitor, collect and aggregate the electricity consumption of the home CEs in a near-real-time way. However, such processes seriously invade consumer privacy, e.g., household occupancy and economic status. Many existing privacy-preserving schemes rely on the assumption that certain components can be trusted, which is hard to be ensured, and cannot resist against collusion attacks. In this study, we leverage differential privacy and a re-encryption model to produce a privacy-preserving aggregation mechanism, called PREEN, for smart-metering communications. PREEN simultaneously facilitates data aggregation, differential privacy, and protection against various attacks. Specifically, we adopt ElGamal algorithm to construct the re-encryption scheme and differential privacy to aggregate the consumption measurements of residential users at the gateways and control center, thereby enabling resistance against slashing privacy attacks. Strict and detailed security proofs as well as extensive performance evaluations are adopted to demonstrate that the proposed PREEN provides rigorous privacy preservation and confidentiality against diverse external and internal attacks, while remaining a comparable or even lower computational overhead and a negligible communication overhead and a reasonable accuracy in terms of the utility of differential privacy, compared to the state-of-the-art mechanisms.

Understanding the intricacies of microbial biofilm formation and its endurance in chronic infections: a key to advancing biofilm-targeted therapeutic strategies.

Bacterial biofilms can adhere to various surfaces in the environment with human beings being no exception. Enclosed in a self-secreted matrix which contains extracellular polymeric substances, biofilms are intricate communities of bacteria that play a significant role across various sectors and raise concerns for public health, medicine and industries. These complex structures allow free-floating planktonic cells to adopt multicellular mode of growth which leads to persistent infections. This is of great concern as biofilms can withstand external attacks which include antibiotics and immune responses. A more comprehensive and innovative approach to therapy is needed in view of the increasing issue of bacterial resistance brought on by the overuse of conventional antimicrobial medications. Thus, to oppose the challenges posed by biofilm-related infections, innovative therapeutic strategies are being explored which include targeting extracellular polymeric substances, quorum sensing, and persister cells. Biofilm-responsive nanoparticles show promising results by improving drug delivery and reducing the side effects. This review comprehensively examines the factors influencing biofilm formation, host immune defence mechanisms, infections caused by biofilms, diagnostic approaches, and biofilm-targeted therapies.

Model exploration of grid adjustment and restoration strategy based on intelligent decision system

In the present scenario, Distributed Energy Resources (DERs) are included into power grid distribution networks to withstand power outages. Restoration is a smart and efficient method for self-healing systems to maintain functionality and re-configure themselves in response to outside breakdowns or attacks. Intelligent decision making framework is required for ensuring that the system as a whole survives external breakdowns and attacks by autonomously reconfiguring itself. Further, the power grid faces serious security concerns, network congestion delay, and packet loss due to improper grid adjustments. As a result, in this research Edge computing assisted Intelligent Decision Model (EC-IDM) has been proposed to address the issue of breakdowns and attacks in real time. The emergence of edge computing has substantially improved the smart grid's ability to operate effectively and communicate effectively by reducing the effects of security concerns, packet loss, latency, and overloaded networks. The integration Smart Grid's peripheral index (SSGPI) optimizes the system's performance and operating efficiency. This research presents a real-time EC-IDM for fixing the electricity grid that can automatically develop the restoration plan and identify the blackout scenario. For restoration choices across several time-and-space axes in EC-IDM, an improved hierarchical coordination technique and a local object optimization method are integrated to optimize breakdowns and attacks in real time. EC-IDM utilizes Smart grid self-healing using distributed automation (DA) to address issues of dependability and reliability of power. The simulation results demonstrates that the suggested approach achieves best in security concerns, optimizes network congestion delay, and packet loss security, improves efficiency and performance compared to the other conventional methods.

Long short-term memory based intelligent control for switched system with a resilient event-triggered communication

This paper studies the event-triggered control for switched systems (SSs) subject to external disturbance and DoS attacks. The adopted long short-term memory (LSTM) neural network with a special gate unit structure has the ability to selectively forget and retain data information. It is used to approximate the external disturbance, and by this offline-trained network approximation, the controller is compensated to eliminate the disturbance. The utilized resilient event-triggering strategy (RETS) can ease network communication burden and reduce the impact of DoS attacks. Further, Zeno phenomenon is avoided by calculating the positive lower limit of time intervals between successive triggers. Moreover, a unified expression that combines switched subsystems subject to DoS attacks is proposed. Then, sufficient conditions for guaranteeing the stability of SSs are derived. Furthermore, the design criteria for controller gains is given using matrix inequalities method. Finally, a simulation example is presented to validate the feasibility and correctness of the proposed method.

Verifiable multi-party quantum secret sharing based on GHZ states

In this paper, we propose a verifiable multi-party secret sharing protocol based on an entanglement swapping theorem of GHZ states. In our protocol, all receivers only need to perform simple GHZ measurements to recover the secret. The quantum-safe system proposed in this paper is feasible with current quantum technologies, and we show that it is secure against both external and internal attacks. In addition, the secret sender can verify whether the secret is correctly recovered.

A Parallel Privacy-Preserving k-Means Clustering Algorithm for Encrypted Databases in Cloud Computing

With the development of cloud computing, interest in database outsourcing has recently increased. However, when the database is outsourced, there is a problem in that the information of the data owner is exposed to internal and external attackers. Therefore, in this paper, we propose decimal-based encryption operation protocols that support privacy preservation. The proposed protocols improve the operational efficiency compared with binary-based encryption operation protocols by eliminating the need for repetitive operations based on bit length. In addition, we propose a privacy-preserving k-means clustering algorithm using decimal-based encryption operation protocols. The proposed k-means clustering algorithm utilizes efficient decimal-based protocols that enhance the efficiency of the encryption operations. To provide high query processing performance, we also propose a parallel k-means clustering algorithm that supports thread-based parallel processing by using a random value pool. Meanwhile, a security analysis of both the proposed k-means clustering algorithm and the proposed parallel algorithm was performed to prove their data protection, query protection, and access pattern protection capabilities. Through our performance analysis, the proposed k-means clustering algorithm shows about 10~13 times better performance compared with the existing algorithms.

Complex Network-Based Resilience Capability Assessment for a Combat System of Systems

With the increasing complexity and frequency of interactions among a large number of heterogeneous nodes within a combat system of systems (SoS), evaluating the capability of the SoS to withstand external attacks and interferences has become an increasingly challenging and urgent issue. However, the complexity of a combat SoS often brings about difficulties in quantitative analysis. This paper proposes a method to assess the resilience of a combat SoS. Firstly, a network model of a combat SoS is constructed, taking into account multi-functional composite nodes. Secondly, the combat capability model of the combat SoS is built based on the capabilities of the kill chains and kill networks. Lastly, an approach is proposed to calculate the combat capability of the system based on the network’s two-terminal connectivity. Moreover, based on the SoS’s combat capability, the resilience is analyzed from the dimensions of anti-destruction, survival, and recovery. The simulation experiments show that, compared to other methods, the method in this paper for calculating the combat capability of the combat SoS does not depend on the number of kill chains and aligns with the combat processes and actual patterns. Additionally, the resilience measurement method can effectively analyze its capability to cope with external interferences.

A MEASUREMENT APPROACH FOR INLINE INTRUSION DETECTION FOR HEARTBLEED-LIKE ATTACKS IN IOT FRAMEWORKS

Cyber security is one of the most crucial aspects of the Internet of Things (IoT). Among the possible threats, great interest is today paid toward the possible capturing of information caused by external attacks on both client and server sides. Whatever the IoT application, the involved nodes are exposed to cyber attacks mainly through the vulnerability of either the sensor nodes themselves (if they have the capabilities for networking operatively) or the IoT gateways, which are the devices able to create the link between the local nodes of the IoT network, and the wide area networks. Due to the low-cost constraints typical of many IoT applications, the IoT sensor nodes and IoT gateways are often developed on low performance processing units, in many cases customized for the specific application, and thus not easy to update against new cyber threats that are continuously identified. In the framework of cyber attacks aimed at capturing sensitive information, one of the most known was the heart bleed, which, has allowed attackers to remotely read protected memory from an estimated 24–55 percent of popular HTTPS sites. To overcome such a problem, which was due to a bug of the OpenSSL, a suitable patch was quickly released, thus allowing to avoid the problem in most of the cases. However, IoT devices may require more advanced mitigation techniques, because they are sometimes unable to be patched for several practical reasons. In this scenario, the paper proposes a novel measurement method for inline detecting intrusions due to heart bleed and heart bleed-like attacks. The proposed solution is based on an effective rule which does not require decoding the payload and that can be implemented on a low-performance general-purpose processing unit. Therefore, it can be straightforwardly implemented and included in either IoT sensor nodes or IoT gateways. The realized system has been tested and validated on a number of experiments carried out on a real network, showing performance comparable (in some cases better) with the heavier machine learningbased methods.

ESUAV-NI: Endogenous Security Framework for UAV Perception System Based on Neural Immunity

Unmanned aerial vehicles (UAV) represent an essential component of advanced intelligent equipment that can be used as an aerial perception system by installing various sensors such as vision, hearing, touch, taste and smell to achieve intelligently integrated perception of environments. However, these perception system with environmental information may be threatened by various internal and external attacks, causing a great challenge to the security of UAV. The original security system relied on an expert knowledge base to prevent attacks, but the weaknesses of lacking proactivity and flexibility are gradually exposed. The strong resistance and survivability of biological systems can be used to fill this capability gap and provide new ideas for the security of UAV perception system. Therefore, an endogenous security framework (ESUAV-NI) based on Neural System (NS) and Immune System (IS) is proposed in this study. Through breeding artificial intelligence (AI) vaccines and distributed neural hierarchical control, achieving security protection for UAV perception system. Moreover, we evaluated the AI vaccine breeding approach in the ESUAV-NI by conducting extensive experiments on internal threats and external aerial imagery camouflage data, respectively. The results show that the proposed approach has superior performance for UAV perception system.

Cloud-based mathematical models for self-organizing swarms of UAVs: design and analysis

Unmanned aerial vehicle (UAV) swarms have gained significant attention for their potential applications in various fields. The effective coordination and control of UAV swarms require the development of robust mathematical models that can capture their complex dynamics. The paper introduces mathematical models and relevant paradigms based on the design and analysis of self-organizing swarms of UAVs. The logical and technological construction of the model relies on the theorems developed by authors for obtaining full information exchange during the swarm quasi-random walk. The suggested rotor-router model interprets the discrete-time walk accompanied by the deterministic evolution of configurations of rotors randomly placed on the vertices of the swarm graph. The recommended optimal and fault-tolerant gossip/broadcast schemes support the resilience of swarm to internal failures and external attacks, and cryptographic protocols approve the security. The proposed cloud network topology serves as the implementation framework for the model, encompassing various connectivity options to ensure the expected behavior of the UAV swarms.

On Cooperative Federated Defense to Secure Multi-Access Edge Computing

The research of cyber security for Multi-access Edge Computing (MEC) has not yet received a great interest. Specifically, the attack detection issue is considered as a major concern since the MEC network, which handles attractive information, is prone to several internal and external network attacks. Recently, the Federated Learning (FL) and Generative Adversarial Network (GAN) have been used to detect and prevent attacks from targeting the wireless mobile networks. In this article, we present the state of the art of MEC attack detection and defense frameworks which incorporate FL and GAN algorithms. Moreover, we propose a new cyber defense framework based on a Federated Generative Adversarial Network (FedGAN) algorithm and non-cooperative game to improve over time the precision of attack detection.

ACCESS CONTROL BY SIGNATURE-KEYS TO PROVIDE PRIVACY FOR CLOUD

Privacy of data in subjects of cloud computing or big data is one of the most principal issues. The privacy methods studied in previous research showed that privacy infringement for cloud computing or big data happened because multi risks on data by external or internal attackers. An important risk to take into consideration when speaking of the privacy of the stored transactions is represented by the transactions’ information which is not in the owner’s control. Such a case is represented by the cloud servers that are administered by cloud providers which cannot be wholly trusted by the users with sensitive, private data such as business plans or private information. A simple method for protecting data privacy is by applying certain privacy techniques onto transactions’ data, followed by the upload of the modified data into the cloud. In this paper, we are proposing a case study that is built on levels containing three models: cloud’s architecture, transaction’s manager and clients. Moreover, we consider that our case study is based on the premise of zero trust among the three models, therefore all the transactions take place with third-parties and the data movements are realized going through various levels of security.

A Location-Privacy-Aware Taxi-Hailing System: Adaptive Differential Privacy-Based Dynamic Incentive Method

Nowadays location-based service (LBS) has become an important service in people’s daily life. Online taxi-hailing system (DiDi, Uber, etc.) is one of the most common LBS system. As the scale of online taxi services continues to expand, some problems have gradually emerged. Specifically, the uncertainty of taxis and passengers makes it difficult to match them effectively, and the passengers’ location privacy will be threatened from both internal and external of the system. In this paper, we first proposed a Bayesian-based location privacy inference attack method from external adversary’s point of view. After that, an adaptive differential privacy-based dynamic incentive method was proposed. Firstly, an adaptivity clocking area division method was proposed to resist the internal privacy threat. Secondly, a dynamic incentive bidding method was proposed to deal with the trading issue. Thirdly, an exponential-based matching method was proposed to resist the external inference privacy threat. Further, theoretical proofs show that the proposed method satisfy the privacy properties of k-anonymity, 2μ1ε-differential privacy, and the economic properties of incentive compatibility, individual rationality. Finally, the experimental results show that the inference attack would achieve a maximum attack success rate of 95% while ensuring the accuracy within 150 meters, and the proposed adaptive differential privacy-based dynamic incentive method can not only provide a good economic performance in terms of satisfaction ratio, social welfare and travel distance, but also can resist internal privacy threat with less than 1% of privacy leakage probability and reduce the success rate of external privacy inference attacks to 25%.

Efficient Detection of Cooperative External Attacks in Wireless Localization Systems

Efficient Detection of Cooperative External Attacks in Wireless Localization Systems

Covert Model Poisoning Against Federated Learning: Algorithm Design and Optimization

Federated learning (FL), as a type of distributed machine learning, is vulnerable to external attacks during parameter transmissions between learning agents and a model aggregator. In particular, malicious participant clients in FL can purposefully craft their uploaded model parameters to manipulate system outputs, which is know as a model poisoning (MP) attack. In this paper, we propose effective MP algorithms to attack the classical defensive aggregation Krum at the aggregator. The proposed algorithms are designed to evade detection, i.e., covert MP (CMP). Specifically, we first formulate the MP as an optimization problem by minimizing the Euclidean distance between the manipulated model and designated one, constrained by Krum. Then, we develop CMP algorithms against the Krum based on the solutions of this optimization problem. Furthermore, to reduce the optimization complexity, we propose low complexity CMP algorithms having only a slight performance degradation. Our experimental results demonstrate that the proposed CMP algorithms are effective and can substantially outperform existing attack mechanisms, such as Arjun's attack and the label flipping attack. More specifically, our original CMP can achieve a high rate of the attacker's accuracy ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\approx 90\%$</tex-math></inline-formula> ). For example, in our experiments using the MNIST dataset, the proposed CMP attacking algorithm against Krum can successfully manipulate the aggregated model to incorrectly classify a given digit as a different one (e.g., 9 as 8). Meanwhile, our CMP algorithm with an approximated constraint can achieve a rate of 87% in terms of the attacker's accuracy (attacker-desired results), with a 73% complexity reduction compared to the original CMP.

Security Issues and Defensive Approaches in Deep Learning Frameworks

Abstract: Deep learning frameworks are instrumental in advancing artificial intelligence, showcasing vast potential across diverse applications. Despite their transformative impact, security concerns pose significant risks, impeding widespread adoption. Malicious internal or external attacks on these frameworks can have far-reaching consequences for society. Our research delves into the intricacies of deep learning algorithms, conducting a thorough analysis of vulnerabilities and potential attacks. To address these challenges, we propose a comprehensive classification system for security issues and corresponding defensive approaches within deep learning frameworks. By establishing clear connections between specific attacks and their defenses, we aim to enhance the robustness of these frameworks. In addition to theoretical considerations, our study extends to real-world applications, exploring a case involving the practical implications of deep learning security issues. Looking ahead, we discuss future directions and open issues within the realm of deep learning frameworks, aspiring to inspire further developments. We anticipate that our research will not only contribute valuable insights but also attract attention from both academic and industrial domains, fostering a collective commitment to fortifying the security of deep learning frameworks.