Target Of Attacks Research Articles

Backdoor Attack Against Dataset Distillation in Natural Language Processing

Dataset distillation has become an important technique for enhancing the efficiency of data when training machine learning models. It finds extensive applications across various fields, including computer vision (CV) and natural language processing (NLP). However, it essentially consists of a deep neural network (DNN) model, which remain susceptible to security and privacy vulnerabilities (e.g., backdoor attacks). Existing studies have primarily focused on optimizing the balance between computational efficiency and model performance, overlooking the accompanying security and privacy risks. This study presents the first backdoor attack targeting NLP models trained on distilled datasets. We introduce malicious triggers into synthetic data during the distillation phase to execute a backdoor attack on downstream models trained with these data. We employ several widely used datasets to assess how different architectures and dataset distillation techniques withstand our attack. The experimental findings reveal that the attack achieves strong performance with a high (above 0.9 and up to 1.0) attack success rate (ASR) in most cases. For backdoor attacks, high attack performance often comes at the cost of reduced model utility. Our attack maintains high ASR while maximizing the preservation of downstream model utility, as evidenced by results showing that the clean test accuracy (CTA) of the backdoored model is very close to that of the clean model. Additionally, we performed comprehensive ablation studies to identify key factors affecting attack performance. We tested our attack method against five defense strategies, including NAD, Neural Cleanse, ONION, SCPD, and RAP. The experimental results show that these defense methods are unable to reduce the attack success rate without compromising the model’s performance on normal tasks. Therefore, these methods cannot effectively defend against our attack.

Cybersecurity Challenges in UAV Systems: IEMI Attacks Targeting Inertial Measurement Units

The rapid expansion in unmanned aerial vehicles (UAVs) across various sectors, such as surveillance, agriculture, disaster management, and infrastructure inspection, highlights the growing need for robust navigation systems. However, this growth also exposes critical vulnerabilities, particularly in UAV package delivery operations, where intentional electromagnetic interference (IEMI) poses significant security and safety threats. This paper addresses IEMI attacks targeting inertial measurement units (IMUs) in UAVs, focusing on their susceptibility to medium-power electromagnetic interference. Our approach combines a comprehensive literature review and QuickField simulation with experimental validation using a commercially available 6-degree-of-freedom (DOF) IMU sensor. We propose a hardware-based electromagnetic shielding solution using mu-metal to mitigate IEMI’s impact on sensor performance. The study combines experimental testing with simulations to evaluate the shielding effectiveness under controlled conditions. The results of the measurements showed that medium-power IEMI significantly distorted IMU sensor readings, but our proposed shielding method effectively reduces the impact, improving sensor reliability. We demonstrate the mechanisms by which medium-power IEMI disrupts sensor operation, offering insights for future research directions. These findings also highlight the importance of integrating hardware-based shielding solutions to safeguard UAV systems against electromagnetic threats.

EAOS: Exposing attacks in smart contracts through analyzing opcode sequences with operands

Today, Ethereum is the world’s largest open-source platform. However, because smart contracts hold a large amount of money and cannot be changed once on the chain, they have become the target of attackers. Users will undoubtedly suffer significant financial losses. To counter these attacks, various methods have been proposed to scan smart contracts for vulnerabilities before deploying them on the blockchain, but few of them determine smart contracts to be vulnerable by examining transactions. In this paper, we propose a framework called EAOS to detect attacks through analyzing the opcode sequences executed by EVM. We first obtain the opcode sequences with operands of smart contracts during EVM execution by replaying the historical transactions of Ethereum, and then extract the feature opcodes from the opcode sequences to generate the feature opcode sequences. Next, we provide some very useful APIs to make it easier for users to get the data related to the opcode sequences. Based on the APIs, users can develop various algorithms to detect attacks. Finally, to verify the effectiveness of EAOS, five algorithms are developed to analyze the replayed transaction opcode sequences. The extensive experimental results demonstrate the effectiveness and efficiency of EAOS and our detection algorithms.

Identification of Security Scenarios in offshore Oil&Gas production facilities based on past incident analysis

Offshore Oil&Gas installations may be attractive targets of intentional attacks due to their presence in socio-political sensitive areas, their economic significance, and the high impact of scenarios that can be triggered by the mismanagement of the hazardous substances handled (e.g., crude oil, natural gas, etc.). In the present study, a new dataset including 112 past security-related incidents that affected the offshore Oil&Gas sector in the last decades was developed and analyzed using Exploratory Data Analysis (EDA), Root Cause Analysis (RCA), and Attack Tree Analysis (ATA). Significant differences were found in the types of threats and attack modes, with terrorism identified as the primary threat, followed by robbery and kidnapping. The identified scenarios caused fatalities, injuries, and asset damages at the targeted facilities, underscoring their potential to cause severe impacts, comparable to the outcomes of safety-related accidents. Tables linking identified attack modes, scenarios, and consequences specific contribution to the SVA workflow proposed by the API Recommended Practices 70 and 70I were obtained. Overall, the results obtained support the application of suitable Security Vulnerability/Risk Assessment (SVA/SRA) methodologies concerning the identification and characterization of adversaries, the definition of attack modes, the evaluation of effectiveness of existing security barriers, and the assessment of consequences.

A hierarchical and secure approach for automotive firmware upgrades

With the development of intelligent and connected vehicles, the expansion of software necessitates an increased significance and frequency of automotive firmware upgrades. The abundance of potential attack vectors and valuable data renders these upgrades enticing targets for attackers. However, the prevailing security services used for automotive firmware upgrades are no longer sufficient to meet security requirements. Hence, this paper proposes a Secure Automotive Firmware Upgrade Approach (SAFUA), aimed at enhancing authentication and communication security during automotive firmware upgrades. To address the heterogeneous performance of in-vehicle nodes and diverse application contexts, this approach introduces multiple authentication modes tailored to various upgrade scenarios. Moreover, hierarchical authentication and secure communication strategies are designed to achieve a balance between security and efficiency requirements. Consolidating these methodologies, a standardized automotive firmware upgrade process is delineated. Formal and informal verification of the proposed approach is conducted to attest its security efficacy. Furthermore, a simulated vehicular environment is constructed to evaluate the temporal and spatial efficiency of the approach across diverse bus and device configurations. The results confirm the adaptability of the secure upgrade approach outlined herein to the automotive firmware upgrade landscape, offering robust security alongside enhanced upgrade efficiency.

A testing and data processing method of projectile explosion position based on three UAVs’ visual spatial constrain mechanism

To accurately obtain the position of the projectile explosion relative to the attack target, this paper propose a new method for measuring the explosive position of projectile using three UAVs’ visual spatial constrain, this method is based on the platform of the UAV equipped with a high-speed camera, and uses the UAV to control the imaging angle of the high-speed camera to form the visual rendezvous mechanism of multiple high-speed cameras. Based on obtaining the pixel coordinate position of projectile explosion in each high-speed camera, we establish the measurement geometric calculation model of the projectile explosive position, and give the coordinates calculation function of projectile explosive that relative to the attack target. To determine the projectile explosion real position, we put forward a measurement data fusion processing method to obtain the real projectile explosion position, and give the data fusion processing algorithm based on projectile explosion position data under the pairwise UAVs visual rendezvous. we also established the engineering uncertainty error calculation model of the test system. Under the testing range of 150 m × 150 m × 30 m, we verify the feasibility and correctness of the proposed test method, the results show that the method proposed in this paper is not only suitable for three-direction coordinate measurement of projectile explosive in the high altitude, but also its measurement error is less than 0.3 m compared with multi-screen optical method.

Killing Two Birds with One Stone: Malicious Package Detection in NPM and PyPI using a Single Model of Malicious Behavior Sequence

Open-source software (OSS) supply chain enlarges the attack surface of a software system, which makes package registries attractive targets for attacks. Recently, multiple package registries have received intensified attacks with malicious packages. Of those package registries, NPM and PyPI are two of the most severe victims. Existing malicious package detectors are developed with features from a list of packages of the same ecosystem and deployed within the same ecosystem exclusively, which is infeasible to utilize the knowledge of a new malicious NPM package detected recently to detect the new malicious package in PyPI. Moreover, existing detectors lack support to model malicious behavior of OSS packages in a sequential way To address the two limitations, we propose a single detection model using malicious behavior sequence, named Cerebro , to detect malicious packages in NPM and PyPI. We curate a feature set based on a high-level abstraction of malicious behavior to enable multi-lingual knowledge fusing. We organize extracted features into a behavior sequence to model sequential malicious behavior. We fine-tune the pre-trained language model to understand the semantics of malicious behavior. Extensive evaluation has demonstrated the effectiveness of Cerebro over the state-of-the-art as well as the practically acceptable efficiency. Cerebro has detected 683 and 799 new malicious packages in PyPI and NPM, and received 707 thank letters from the official PyPI and NPM teams.

A novel optimized deep learning based intrusion detection framework for an IoT networks

The burgeoning importance of Internet of Things (IoT) and its diverse applications have sparked significant interest in study circles. The inherent diversity within IoT networks renders them suitable for a myriad of real-time applications, firmly embedding them into the fabric of daily life. While IoT devices streamline various activities, their susceptibility to security threats is a glaring concern. Current inadequacies in security measures render IoT networks vulnerable, presenting an enticing target for attackers. This study suggests a novel dealing to address this challenge through the execution of Intrusion Detection Systems (IDS) leveraging superior deep learning models. Inspired by the benefits of Long Short Term Memory (LSTM), we introduce the Genetic Bee LSTM(GBLSTM) networks for the development of intelligent IDS capable of detecting a wide range of cyber-attacks targeting IoT area. The methodology comprises four key execution: (i) collection of unit for profiling normal IoT device behavior, (ii) Identification of malicious devices during an attack, (iii) Prediction of attack types implemented in the network. Intensive experimentations of the suggested IDS are conducted using various validation methods and prominent metrics across different IoT threat scenarios. Moreover, comprehensive experiments are conducted to evaluate the suggested models alongside existing learning models. The results demonstrate that the GBLSTM-models outperform other intellectual models in terms of accuracy, precision, and recall, underscoring their efficacy in securing IoT networks.

An Efficient Flow-Based Anomaly Detection System for Enhanced Security in IoT Networks.

The growing integration of Internet of Things (IoT) devices into various sectors like healthcare, transportation, and agriculture has dramatically increased their presence in everyday life. However, this rapid expansion has exposed new vulnerabilities within computer networks, creating security challenges. These IoT devices, often limited by their hardware constraints, lack advanced security features, making them easy targets for attackers and compromising overall network integrity. To counteract these security issues, Behavioral-based Intrusion Detection Systems (IDS) have been proposed as a potential solution for safeguarding IoT networks. While Behavioral-based IDS have demonstrated their ability to detect threats effectively, they encounter practical challenges due to their reliance on pre-labeled data and the heavy computational power they require, limiting their practical deployment. This research introduces the IoT-FIDS (Flow-based Intrusion Detection System for IoT), a lightweight and efficient anomaly detection framework tailored for IoT environments. Instead of employing traditional machine learning techniques, the IoT-FIDS focuses on identifying unusual behaviors by examining flow-based representations that capture standard device communication patterns, services used, and packet header details. By analyzing only benign traffic, this network-based IDS offers a streamlined and practical approach to securing IoT networks. Our experimental results reveal that the IoT-FIDS can accurately detect most abnormal traffic patterns with minimal false positives, making it a feasible security solution for real-world IoT implementations.

Virtuous victimhood as a Dark Triad resource transfer strategy

Virtuous victim signals have been described as a resource transfer strategy motivated by dark traits of narcissism and Machiavellianism Ok et al. (2021). Here we report direct replication of key predictions of this claim, test robustness to alternative predictors, and explore possible roles for sadism within virtuous-victim signalling. Study 1 (N = 750, preregistered) replicated significant, large, associations of virtuous victim signalling with communal narcissism and Machiavellianism. Study 2 (N = 750, preregistered) tested robustness of the association using an alternative victim signalling measure. The results again replicated, this time with larger effects (β = 0.41 for narcissism and β = 0.22 for Machiavellianism). Virtuous-victim status releases resources to the claimant but also leaves the accused as targets for attack. We therefore explored (study 3) whether a further dark trait – sadism –predicted exploitation of this victimization opportunity. Sadism was significantly and specifically associated with engaging in and enjoying attacks on accused. The results provide independent support for virtuous victim signalling as narcissistic Machiavellianism. Dark traits may be adapted to exploit the resources released to victims, and opportunities for resource-capture provided by victimizing others under a righteous guise.

Three‐body confrontation differential game for an aggressive attacker

AbstractIn this work, we study the three‐body confrontation differential game, in which an Attacker attempts to capture a Target while a Defender guards the latter. We decompose the three‐body game into two subgames based on the game of kind theory: the included angle game and the distance ratio game. By integrating the Apollonius circle with these subgames, we define the semipermeable surface and infer the optimal guidance strategies for the three players on this surface. Unlike existing methods, our method effectively addresses the coupling relationship between the target, defender, and attacker, resulting in greater applicability. Based on the derived semipermeable surface, the conditions for the three players to win the game are also presented. These winning conditions are comprehensively considered from both lines of sight, including the angle and the ratio between the Attacker–Target distance and the Attacker–Defender distance, thereby providing a game solution through an analysis of the players' dynamics. From the perspective of viewing angle ratio, our method bridges the gap in the viewing angle game outcomes compared to existing methods. To demonstrate the effectiveness of our method, we provide simulation examples that illustrate the three‐body confrontation game in various scenarios.

Backdoor Attacks and Defenses Targeting Multi-Domain AI Models: A Comprehensive Review

Since the emergence of security concerns in artificial intelligence (AI), there has been significant attention devoted to the examination of backdoor attacks. Attackers can utilize backdoor attacks to manipulate model predictions, leading to significant potential harm. However, current research on backdoor attacks and defenses in both theoretical and practical fields still has many shortcomings. To systematically analyze these shortcomings and address the lack of comprehensive reviews, this paper presents a comprehensive and systematic summary of both backdoor attacks and defenses targeting multi-domain AI models. Simultaneously, based on the design principles and shared characteristics of triggers in different domains and the implementation stages of backdoor defense, this paper proposes a new classification method for backdoor attacks and defenses. We use this method to extensively review backdoor attacks in the fields of computer vision and natural language processing, and also examine the current applications of backdoor attacks in audio recognition, video action recognition, multimodal tasks, time series tasks, generative learning, and reinforcement learning, while critically analyzing the open problems of various backdoor attack techniques and defense strategies. Finally, this paper builds upon the analysis of the current state of AI security to further explore potential future research directions for backdoor attacks and defenses.

Mitochondrial reactive oxygen species-mediated fibroblast activation has a role in tumor microenvironment formation in radiation carcinogenesis.

Cancer risks attributable to low-dose and low-dose-rate radiation are a serious concern for public health. Radiation risk assessment is based on lifespan studies among Hiroshima-Nagasaki A-bomb survivors; however, there are statistical limitations due to a small sample size for low-dose radiation. Therefore, basic biological studies are helpful in understanding the mechanism of radiation carcinogenesis. The detrimental effects of ionising radiation (IR) are caused by reactive oxygen species (ROS)-mediated oxidative DNA damage. IR-induced delayed ROS are produced in the electron transport chain reaction of the mitochondrial complex. Thus, mitochondria are a source of ROS and a primary target for ROS attacks. Consequently, mitochondrial dysfunction is thought to be a key event in the metabolic changes of cancer cells and is important in radiation-induced carcinogenesis. In this paper, we present recent findings on radiation carcinogenesis effect assessment, focusing on mitochondrial function as stress sensors.

Pattern Analysis of Cyber Crime Incidents to Predict Occurrence and Selection of The Best Technology to Prevent IT

Cybercrime has become a significant global threat, affecting individuals, organizations, and governments alike. This research paper focuses on the pattern analysis of cyber crime incidents and aims to develop a predictive model to forecast its occurrence while selecting the most effective technology for prevention. Through an extensive literature review, we analyze existing research on cybercrime patterns, predictive modeling, and technologies for prevention. The study identifies the gaps in the current literature, emphasizing the need for a comprehensive approach that combines data-driven analysis with advanced cybersecurity technologies. The research methodology encompasses data collection from diverse sources, followed by rigorous data preprocessing and cleaning to ensure data quality. Machine learning algorithms and statistical methods are utilized for developing the predictive model. Evaluation metrics are employed to measure the model's performance and its ability to predict cybercrime incidents accurately. The pattern analysis of cybercrime incidents reveals various attack vectors, target sectors, and geographical distributions, providing crucial insights into the modus operandi of cybercriminals. This analysis forms the basis for building the predictive model, which can assist law enforcement agencies and cybersecurity professionals in anticipating and preventing cybercrime effectively. The findings of this research contribute significantly to the field of cybercrime prevention. The developed predictive model enhances early warning capabilities, enabling proactive measures against cyber threats. Additionally, the technology selection framework assists organizations in making informed decisions regarding cybersecurity investments.

Biophysical Analysis of Vip3Aa Toxin Mutants Before and After Activation.

Cry toxins from Bacillus thuringiensis are effective biopesticides that kill lepidopteran pests, replacing chemical pesticides that indiscriminately attack both target and non-target organisms. However, resistance in susceptible pests is an emerging problem. B. thuringiensis also produces vegetative insecticidal protein (Vip3A), which can kill insect targets in the same group as Cry toxins but using different host receptors, making the combined application of Cry and Vip3A an exciting possibility. Vip3A toxicity requires the formation of a homotetramer. Hence, screening of Vip3A mutants for increased stability requires orthogonal biophysical assays that can test both tetrameric integrity and monomeric robustness. For this purpose, we have used herein for the first time a combination of analytical ultracentrifugation (AUC), mass photometry (MP), differential static light scattering (DSLS) and differential scanning fluorimetry (DSF) to test five mutants at domains I and II. Although all mutants appeared more stable than the wild type (WT) in DSLS, mutants that showed more dissociation into dimers in MP and AUC experiments also showed earlier thermal unfolding by DSF at domains IV-V. All of the mutants were less toxic than the WT, but toxicity was highest for domain II mutations N242C and F229Y. Activation of the protoxin was complete and resulted in a form with a lower sedimentation coefficient. Future high-resolution structural data may lead to a deeper understanding of the increased stability that will help with rational design while retaining native toxicity.

A Survey of Text Watermarking in the Era of Large Language Models

Text watermarking algorithms are crucial for protecting the copyright of textual content. Historically, their capabilities and application scenarios were limited. However, recent advancements in large language models (LLMs) have revolutionized these techniques. LLMs not only enhance text watermarking algorithms with their advanced abilities but also create a need for employing these algorithms to protect their own copyrights or prevent potential misuse. This work conducts a comprehensive survey of the current state of text watermarking technology, covering four main aspects: (1) an overview and comparison of different text watermarking techniques; (2) evaluation methods for text watermarking algorithms, including their detectability, impact on text or LLM quality, and robustness under target or untargeted attacks; (3) potential application scenarios for text watermarking technology; and (4) current challenges and future directions for text watermarking. This survey aims to provide researchers with a thorough understanding of text watermarking technology in the era of LLMs, thereby promoting its further advancement.

A context-aware zero trust-based hybrid approach to IoT-based self-driving vehicles security

With the speedy progression and adoption of IoT devices in modern self-driving vehicles (SDVs), autonomous driving industry is gradually reforming its capabilities to provide better transportation services. However, this domain faces enormous security and privacy challenges and thus has become an attractive target for attackers due to its rapid growth and market worth. Furthermore, the rapid transformation in technological tools in transport industry and speedy evolution of cyber-attacks paved the way for designing efficient IDSs. Motivated by these challenges, we put forward a new secure and efficient IDS approach for the security of SDVs. The propose approach utilizes an emerging strategy to mitigate security vulnerabilities and cyber attacks detection using zero trust (ZT) model. Through this work, we put forward a context-aware zero trust security framework for IoT-based SDVs. The proposed framework utilizes a context-aware design to evaluate the trustworthiness of the devices using multi-source trust and reputation model. Then, to make the framework more effective and reliable, we introduce crawler system into the context of the IoT-devices in SDVs to make the system unbiased. Additionally, an observer module is developed that employs state-of-the-art machine learning algorithm to detect malicious actions. Empirical results on two standard benchmark datasets (i.e., Car Hacking and ToN_IoT) validate the practicality and robustness of propose framework in real-world transport systems with enhanced security and trust management against evolving cyber-threats. Detection results demonstrate that the proposed framework secured the best performance by achieving 99.43% and 99.52% accuracy for Car Hacking and ToN_IoT, respectively. The findings of this study will help the security professionals and researchers to comprehend the importance of ZT architecture in developing effective and robust security solutions for modern IoT-based SDVs.

The Impact of Terrorism on Tourism Development in the Republic of Kosovo

In the last twenty years, tourism has faced many terrorist attacks or similar crises, so we can ask the question of what is the connection between terrorism and tourism, how does all this affect global tourism. The targets of terrorist attacks in a certain tourist destination can vary, starting from small objects to complex tourist infrastructure, and the basis of the attacks can be political, economic, religious, or other reasons. This research has special theoretical and practical importance for the development of tourism in Kosovo, but from the aspect of the impact of terrorism on tourism, with special reference to the Republic of Kosovo as a tourist destination. Research related to the impact of terrorism on tourism in the Republic of Kosovo will help us to understand, in essence, the negative effects that terrorism can have on the development of tourism in this country. From this research, we are expected to get to know the most important aspects of terrorism and tourism in modern conditions, but also with the legislation in this field and how it can help prevent terrorist acts in tourist facilities in Kosovo.

Leveraging Universal Adversarial Perturbation and Frequency Band Filters Against Face Recognition

Universal adversarial perturbation (UAP) exhibits universality as it is independent of specific images. Although previous investigations have shown that the classification of natural images is susceptible to universal adversarial attacks, the impact of UAP on face recognition has not been fully investigated. Thus, in this paper we assess the vulnerability of face recognition for UAP. We propose FaUAP-FBF, which exploits the frequency domain by learning high, middle, and low band filters as an additional dimension of refining facial UAP. The facial UAP and filters are alternately and repeatedly learned from a training set. Furthermore, we convert non-target attacks to target attacks by customizing a target example, which is an out-of-distribution sample for a training set. Accordingly, non-target and target attacks form a uniform target attack. Finally, the variance of cosine similarity is incorporated into the adversarial loss, thereby enhancing the attacking capability. Extensive experiments on LFW and CASIA-WebFace datasets show that FaUAP-FBF has a higher fooling rate and better objective stealthiness metrics across the evaluated network structures compared to existing universal adversarial attacks, which confirms the effectiveness of the proposed FaUAP-FBF. Our results also imply that UAP poses a real threat for face recognition systems and should be taken seriously when face recognition systems are being designed.

Mechanism of formation and significance of antimitochondrial autoantibodies in the pathogenesis of primary biliary cholangitis

Primary biliary cholangitis (PBC) is a chronic cholestatic progressive liver disease associated with cholangiopathies. The detection of antimitochondrial autoantibodies (AMAs) plays an important role in the diagnosis of classical PBC. AMAs are formed against the antigenic component associated with the dihydrolipoyl transacetylase of pyruvate dehydrogenase complex (E2 PDC) localized on the inner membrane of mitochondria. The loss of immune tolerance of E2 PDC in PBC is thought to be the cause of the mechanism of AMA formation and immune-mediated destruction of biliary epithelial cells (BECs) of the small- and medium-sized intrahepatic bile ducts. E2 PDC is not only present in BECs, but is also abundant in the mitochondria of all nucleated cells. The question remains as to why E2 PDC of only small BECs is the target of autoimmune attack. There is no evidence that AMAs have a deleterious effect on BECs. New scientific data has emerged that explains the damage to BECs in PBC by the defect of the biliary bicarbonate (HCO3–) “umbrella” that protects BECs from the detergent action of bile acids under physiological conditions. Disruption of HCO3– production by BECs in PBC leads to changes in the pH of hepatic bile, accompanied by accumulation of bile acids in the small BECs. The detergent action of bile acids leads to damage of membrane structures of BECs and their apoptosis, development of ductulopenia, and intrahepatic cholestasis. For the first time, it has been suggested that under the influence of bile acids, the E2 PDC antigen may undergo conformational changes that alter its immunological properties. E2 PDC becomes a neoantigen that is recognized by the normal (“healthy”) immune system as a foreign antigen, leading to the production of AMAs. For the first time, the authors of this review provide an explanation for why only small BECs are damaged in PBC.