1,721,094 research outputs found
Image-Based Android Malware Detection Using Deep Learning
No full textThe Android operating system (OS) dominates the mobile phone OS industry, with over 70% of the market share. With the growth of Android OS-based smartphones, it has become a prime target for mobile malware attacks. Minimal alterations in malware samples can easily evade traditional detection methods such as signature-based detection. In contrast, artificial intelligence (AI) and machine learning (ML)-based malware detection has proven more effective, as it can detect zero-day malware. Previous studies have shown that AI/ML-based malware classifiers trained on categorical features are vulnerable to adversarial evasion attacks. Therefore, in this study, we transform the features extracted from Android apps into image-based data and investigate the performance of Convolutional Neural Networks (CNNs), Inception Networks, and Residual Networks (ResNet) on this data. We employ 41,382 Android malware samples belonging to 240 malware families and 36,755 benign apps to train and test the models. Our experiment results show that CNNs outperform Inception Networks and ResNet with up to 99% classification accuracy. Furthermore, our analysis also indicates that CNNs trained on image-based Android malware and benign data outperform various Android malware detection techniques proposed in the literature
The Threat Landscape of Connected Vehicles
No full textAs connected vehicles (CVs) play an increasingly pivotal role in modern transportation, cybersecurity threats targeting these systems have become a critical area of concern. This study systematically identifies and classifies vulnerabilities from the National Vulnerability Database (NVD) and the Automotive Attack Database (AAD) using a semi-automated filtering process. Our analysis identifies a total of 508 vulnerabilities across these databases, which are categorised based on ISO/SAE 21434 impact categories: safety, financial, operational, and privacy. A key finding reveals that 14.6\% of these vulnerabilities have systemic implications, meaning they have the potential to cause widespread disruption across multiple vehicles or the broader transportation network. Furthermore, 45\% of the vulnerabilities are associated with remote attack vectors, significantly increasing the risk of large-scale exploitation. This research contributes an updated database of automotive vulnerabilities, providing a valuable resource for the cybersecurity community. The findings highlight the need to enhance current automotive cybersecurity standards, such as ISO/SAE 21434, to address the complex inter-dependencies and systemic risks within connected vehicle ecosystems
Near Space Instability. Geopolitical Tensions, Debris Crisis, and Cyberattacks
No full textItem is not available in this repository.Stefania Paladini - ORCID: 0000-0002-1526-3589 https://orcid.org/0000-0002-1526-3589Never before Near Space, that portion of outer space closest to Earth and crucial for human activities, has been under threat like in present days. There are several factors responsible for the current status of things, but three of them (geopolitics, debris and cyberthreats) are emerging as the most critical in terms of impact and long-term implications. What is more important, those factors are now colliding, with state-sponsored cyberattacks in outer space that risks exacerbating the debris crisis. The aim of this study is therefore to investigate how geopolitical tensions are pushing existing criticalities such as the debris crisis in the Earth Orbit and the growing threat of cyberattacks to satellite and ground space infrastructure toward an unprecedent level of tensions. Building on the analysis of existing datasets, it will present some cases for discussion and attempts a scenario analysis for the short-medium term.https://doi.org/10.1007/978-3-031-82031-1_3pubpu
Comprehensive Analysis of MSpy within covert operations.
Full text linkThe rapid development of mobile technology is creating new opportunities for crime targeting smart devices and wireless networks, posing major challenges to public safety. Law enforcement must skilfully collect and present digital evidence in court to maintain its reliability and accuracy. The main topics of this research are the characteristics, applications, and effects of mobile spyware in law enforcement and intelligence operations. Through in-depth research, we explore the characteristics of mobile spyware and how they affect the integrity and recovery of forensic evidence. This study examines the latest advances in spyware technology and methods for collecting real-time statistics on calls, messages, images, social media interactions, and location data. Previous research shows that sophisticated mobile spyware often evades detection, making it easier to gather evidence. Such sophisticated spyware can challenge established forensic techniques by removing or altering important information. However, this study shows how effective cell phone surveillance and detective programs like mSpy are in tracking crimes and improving the quality of evidence collection. The results show how using the latest tools can lead to better detection and analysis. In addition, the research emphasizes the importance of approaches to solve the growing challenges of mobile devices. The study then provides law enforcement organisations and covert agents with useful information about the effectiveness of mobile spy apps and programs like mSpy in crime prevention. By using this cutting-edge technology, law enforcement agencies can greatly improve the efficiency of their investigations and ensure the integrity of digital evidence used in court
Towards effective malware clustering : reducing false negatives through feature weighting and the Lp metric
No full textIn this paper we present a novel method to reduce the incidence of false negatives in the clustering of malware detected during drive-by-download attacks. Our method comprises the use of a high-interaction client honey-pot called Capture-HPC to acquire behavioural system and network data, and the application of clustering analysis. Our method addresses various issues in clustering, including (i) finding the number of clusters in the dataset, (ii) finding good initial centroids, (iii) determining the relevance of each of the features at each cluster. Our method applies partitional clustering based on the Minkowski Weighted K-Means (Lp) and anomalous pattern initialization. We have performed various experiments on a dataset containing the behaviour of 17,000 possibly infected websites gathered from sources of malicious URLs. We find that our method produces a smaller within cluster variance and a lower quantity of false negatives than other popular clustering algorithms such as K-Means and the Ward's method
Differentially Private Spiking Neural Networks: Enhancing Privacy and Robustness in Social Robotics.
Full text linkAs social robots increasingly integrate into various sectors,
concerns over privacy and security become paramount. Social robots, while capable of processing and interacting with sensitive user data, often face significant limitations in processing power and energy efficiency. Spiking Neural Networks (SNNs), inspired by biological neurons, provide a promising solution by offering efficient temporal processing with reduced energy consumption. However, similar to conventional Artificial Neural Networks (ANNs), SNNs are vulnerable to privacy attacks, such as model inversion and membership inference, which can expose sensitive training data. We proposes the use of Differential Privacy (DP) to safeguard user data in SNN models for social robots. We train two models
on the MNIST dataset: a baseline SNN and a differentially private SNN (DP-SNN). We evaluate both models through a privacy attacks, demonstrating how differentially private SNNs mitigate data leakage. We assess the ability of DP-SNNs to withstand malicious inputs, showing that the noise introduced by differential privacy enhances robustness in addition to privacy preservation. Our results indicate that differentially private SNNs not only maintain strong privacy guarantees but also improve resilience against adversarial attacks, making them an ideal solution for
social robots where both data security and processing efficiency are critical
AI-Driven Risk Assessments: Advancing Cybersecurity and Sustainability
No full textIn an increasingly interconnected world, the complexity of global security, safety, and sustainability challenges is escalating. Digital transformation, cyber threats, and environmental risks demand dynamic and comprehensive risk management approaches. This paper explores the potential of AI-driven risk assessments in addressing these challenges, focusing on how AI enhances both cybersecurity and sustainability efforts. AI technologies enable real-time threat detection, predictive analytics, and proactive risk mitigation, offering significant improvements over traditional risk management methods. Furthermore, the introduction of quantum computing in AI-driven risk assessments provides unprecedented computational power, enhancing the accuracy and speed of risk identification and mitigation strategies. However, the integration of AI and quantum computing also introduces ethical concerns related to bias, transparency, and accountability. This study critically examines the effectiveness of AI-driven models in various sectors, including cybersecurity, critical infrastructure, and environmental management, highlighting both their advantages and limitations. By addressing emerging risks through the convergence of AI, privacy, and security frameworks, this paper emphasises the need for collaboration between governments, industries, and academia to ensure the ethical and effective application of AI and quantum technologies in risk management. The findings present a pathway for balancing innovation with responsible AI deployment in an increasingly volatile global landscape
Going Beyond Counting First Authors in Author Co-citation Analysis
Full text linkThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
- …
