This study presents a cross-domain cybersecurity strategy that includes a sophisticated intrusion detection system (IDS) positioned strategically at the default gateway of Healthcare Systems and Industrial Internet of Things (IIoT) infrastructures. The suggested model combines a Convolutional Neural Network (CNN) with a Gated Recurrent Unit (GRU) architecture and incorporates a simulated XGBoost layer to improve predicting accuracy. The CNN-GRU model is designed to extract intricate spatial and temporal characteristics. The CNN component focuses on collecting local patterns, while the GRU component is responsible for addressing sequential dependencies. The outputs are merged with the original input features to create a comprehensive collection of features. This set is then passed through dense layers that simulate the capabilities of XGBoost for the final prediction. This ensemble methodology utilizes deep learning and gradient boosting to enhance accuracy and resilience in classification tasks. The framework prioritizes the processing of data in real-time, guaranteeing the integrity of the system and enabling swift detection of threats. The effectiveness of the model can be measured by its packet processing rate, which demonstrates its performance under different workloads. Our Intrusion Detection System (IDS) establishes a higher benchmark for proactive and dynamic defense, offering strong safeguarding capabilities for healthcare, IIoT, and general networks. It continuously adjusts to changing cyber threats in order to ensure the security of key infrastructure.

In recent years, the use of the Internet of Things (IoT) has grown rapidly in smart cities, households, and smart industries environments due to the benefits of convenience and easy implementation. The Long-Range Wide Area Network (LoRaWAN) is a key technology driving this growth, which has seen significant adoption due to its scalability, long-range capabilities, low power consumption, lost cost and wide area coverage. As the use of LoRaWAN-enabled IoT environment has increased, the cyber-attacks related to it have also increased. About 95% of the cyber-attacks in IoT environment occurs due to the security vulnerabilities caused by the user's errors. However, much research in the LoRaWAN environment is highly focused on the security vulnerabilities due to technical errors, leaving gaps in identifying the security vulnerabilities that occur due to user errors. Therefore, this paper addresses these gaps by identifying user-centric errors which may lead to cyber-attacks in an LoRAWAN IoT environment. It builds a taxonomy based on a comprehensive literature review and maps these user errors to potential security vulnerabilities within LoRaWAN-enabled IoT environments.

In the present world, Software-Defined Networks (SDN) is the developing technological environment that allows integrated control and splits the control plane from the data plane. It is essential to classify the attacks in SDN-based networks to improve security. Concomitantly, SDN-related networks are vulnerable to numerous attacks especially Distributed Denial of Service (DDoS) attacks which interrupt the data transmission and network data. To resolve this issue, various traditional researchers have attempted to attain attack detection, however, there is a lack in computational cost, accuracy rate, and Packet Delivery Ratio. To overcome the limitations, the proposed hybrid model employs a system creation model for encrypting data using ELGAMAL and ECC algorithms to strengthen data security. Furthermore, the present hybrid model incorporates the wrapper-based approach Sine Cosine hybrid optimization algorithm with Modified Particle Swarm Optimization (SCMPSO) for feature selection which is intended to improve the performance of the classification. Furthermore, it utilizes the XG Boost-Light Gradient-Boosting Machine (GBM) algorithm for attack classification. Accordingly, Extreme Gradient-Boosting (XG Boost) can handle the missing data, and the ensemble nature of XG Boost with multiple Decision trees (DT) combinations makes it challenging to finalize the prediction. To overcome the limitations of the XG Boost algorithm, the proposed model uses Light GBM. Correspondingly, the present method created a dataset using the Mininet tool. The efficacy of the proposed hybrid model is calculated with several evaluation metrics to analyze the performance. The proposed method is intended to contribute to SDN-based network development and is envisioned to contribute to attack detection mechanisms.

In current times, network slicing in a 5G context is a significant study field. But it might be difficult to meet network slice requests' requirements. Network slices need to share limited resources; therefore energy efficiency and security are crucial. Additionally, it is essential to establish secured network slicing for Software-Defined Network/Network Function Virtualization (SDN/NFV). As attackers have developed to become more skilled and frequently use different attacking approaches, security is a crucial problem in network slicing. We address security, ineffective network slicing, and overloading using load balancing and Deep Learning (DL) based network slicing algorithms in edge enabled SDN/NFV assisted 5G settings in this research. Here, we mainly concentrate on secure and efficient network slicing in SDN/NFV assisted 5G systems. Initially, slicing of network is performed based on UniqueNet which includes lightweight convolutional layers that reduce the processing time and increase accuracy. For authentication of users, we employ the Improved Mersenne Twister (IMT) algorithm and role-based access control is performed using Improved Deep Q Network (ImDQN) algorithm for authorization purpose. Clustering is done by using k-means clustering (KMC) algorithm. Here, Cluster Head (CH) performed intrusion detection using Enhanced Bidirectional Generative Adversarial Network (E-BiGAN) algorithm. After detected intrusions, the Kangaroo-based IDS (KIDS) jump and send the notification to all the nodes in the CH. For efficient load balancing, we perform optimal switch selection using Dove Swarm Optimization (DSO). The performance of the suggested framework is then evaluated in terms of different metrics and compared with existing approaches to prove the efficacy of the proposed system.

Land Cover maps are one of the crucial descriptors in land use planning, management, and development. However, there is a rapid change in land cover due to urbanization and agricultural land development. In this paper, a Residual Attention U-Net Model was trained to classify and map the Land Cover in the Philippines using Satellite Images. The satellite image was given by DOST DATOS - Remote Sensing and Data Science Help Desk which provides free satellite Images for academic use. The image dataset was labeled with eight (8) land cover classes. Data pre-processing and augmentation were used to improve the image and mask dataset. The trained model was evaluated by segmentation metrics namely, Jaccard index and Dice Score. After tedious training, the model achieved a performance of 62% dice coefficient and 45% jaccard index with a focal loss of 16%. This shows that the model can predict the majority of the classes however, there are some classes that are predicted incorrectly which is understandable given the performance of the model. A large-scale satellite image was also fed into the model which automatically produced land cover segmentation.

This paper presents a deep learning approach for drone navigation aimed at accurately estimating a drone's position based on RSS in a simulated environment. A predefined route is established, and the drone collects RSS and IMU data while traversing it multiple times. This dataset serves as the foundation for training various neural network architectures. We create multiple training and test datasets to compare different models and identify the most effective approach for predicting the drone's location. After training, the neural networks are evaluated using distinct test data, showing that deep learning models can reliably predict a drone's position in simulation. The developed predictive models enhance the accuracy and reliability of autonomous drone operations in real-world scenarios.

Bushfire severity prediction plays a vital role in mitigating the catastrophic effects of bushfires on ecosystems, human life, and property. Accurate prediction models enable timely interventions and informed decision-making, which are crucial for effective fire prevention and response strategies. This study aims to develop an efficient bushfire severity prediction model by leveraging satellite-derived data, specifically from Landsat 8, which includes spectral indices, topographical and climatic features. The primary objective of this study is to assess the impact of different feature combinations on model performance using the XGBoost regression model with dNBR as the target variable. Eighteen tests were conducted, ranging from single-feature indices like NDVI to more complex multi-feature sets. The findings revealed that a 21-feature combination achieved the highest accuracy of 90.09%, with a MSE of 0.0008, a computational time of 1201.86 seconds, and a memory usage of 106.28 MB. However, adding more features did not always result in proportional improvements and often led to increased computational demands. In contrast, tests with fewer features demonstrated that high precision could be achieved with reduced resource consumption, offering a favourable trade-off between performance and efficiency. The results underscore the importance of strategic feature selection and optimization to enhance model accuracy while minimizing computational overhead. This study provides valuable insights for creating effective, high-performance fire severity prediction models, especially in resource-constrained settings.

This study looks at how Microsoft AirSim and OpenAI's Natural Language Processing capabilities may be used to enable drone navigation in a campus simulation. The project uses Unreal Engine to create a 3D simulation of Georgia State University's campus to investigate the use of language-based drone control. Three current technologies are combined in our implementation: (1) Microsoft's AirSim platform for simulating drone physics, (2) OpenAI's ChatGPT API for natural language interpretation and command processing, and (3) a comprehensive campus environment integrated into Unreal Engine. This integration replaces traditional drone control interfaces by allowing users to operate simulated drones through natural language instructions. By converting user commands into drone navigation directions, this technology demonstrates the practical applications of language models. Our findings show that this approach can enhance campus navigation simulations and provide a secure environment for testing drone operations in urban areas. Our study highlights the potential of combining language processing with drone control systems, particularly within educational simulations.

Cloud computing has become a significant part of people's daily lives over the last decade due to its cost-effective services, such as Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS). As a result, the cloud environment contains a massive amount of sensitive and confidential data, thus becoming a crucial target for attackers. Additionally, cloud forensics over the cloud environment has become complex and difficult due to the cloud's several challenges. These challenges include collecting volatile data, dependency on cloud service providers (CSPs), secure logs storage, and more. This paper proposes a Cloud-Based Data Volatility and CSP Reliance Eradication Framework (CDCEF) to mitigate data volatility and dependency on CSP in cloud forensics in IaaS. We also performed experiments on the Amazon Web Service (AWS) Lightsail - a widely used cloud platform by people globally, based on hypothetical cybercrime scenarios to support our framework. The significant benefits of this framework for investigators are that it allows them to collect volatile data without losing the integrity of the evidence and eradicates CSP reliance. Additionally, the framework provides another benefit of minimizing the usage of forensic tools.

Cybersecurity remains a critical issue due to the intrinsic vulnerability in the Cyberspace which threatens national borders, affecting economies, governments, and ordinary citizens. Cyberattacks are ever-escalating and becoming more sophisticated and widespread, requiring robust national cybersecurity frameworks to safeguard against these evolving threats. This paper examines the essential role of government in developing, implementing, and strengthening national cybersecurity frameworks. It explores key areas such as regulatory frameworks, public-private partnerships, international cooperation. The paper argues that governments must adopt a proactive, comprehensive approach to cybersecurity, balancing security measures with fundamental rights and freedoms.

5G slices are vulnerable to indirect Distributed Denial of Service (DDoS) attacks, by which a flooded traffic is sent toward a target and makes it unavailable. When a DDoS attack is launched against a single slice, the other slices with shared infrastructure with the target are also affected. Most existing mitigation techniques require a detection phase, which is insufficient for unknown and complex attacks. Moving Target Defense (MTD) is a security mechanism that invalidates the adversary's collected information, and it can be deployed without the detection phase. In this paper, we propose a Slice Mutation technique based on Reinforcement Learning (SMRL) that reduces the impact of DDoS attacks on 5G slices while keeping the number of allocated slices acceptable. SMRL proposes a general RL model that considers ternary and ranking numbers to improve learning performance. We tested a new system, SMRL, on computer networks that were attacked by a real botnet called Mirai. We evaluated the system using different measures, including a new way of looking at how the system works. The reported results show that SMRL reduces the number of slices affected by a DDoS attack and improves the distribution of the slices among infrastructure resources by 46% and 20%, respectively.

Controlling access to data is essential in ensuring data is only accessed by authorised and trusted users. For these reasons, zero-trust frameworks have been in the centre of interest in the past few years. Zero-Trust frameworks assume that users and systems have been compromised and deal with them as untrusted entities that requires multiple levels of authorisation and security attributes to be compliant in order to be considered trusted. The most used zero trust frameworks use static thresholds to grant levels of access to systems which could introduce false positives and incorrect access privileges to systems/networks. This research paper proposes a machine- learning (ML)-based zero-trust framework that utilises an anomaly detection algorithm. The output of the anomalous detection would inform the observers the trustworthiness of systems in their environments. Moreover, performance, complexity and impact of our proposed scheme is compared against a static threshold zero-trust framework.

This study introduces the development of a stateof-the-art, real-time ECG monitoring and analysis system, incorporating cutting-edge medical technology and innovative data security measures. Our system performs three distinct functions: real-time ECG monitoring and disease detection, encrypted storage and synchronized visualization, and statistical analysis on encrypted data. At its core, the system uses a three-lead ECG preamplifier connected through a serial port to capture, display, and record real-time ECG data. These signals are securely stored in the cloud using robust encryption methods. Authorized medical personnel can access and decrypt this data on their computers, with AES encryption ensuring synchronized real-time data tracking and visualization. Furthermore, the system performs statistical operations on the ECG data stored in the cloud without decrypting it, using Fully Homomorphic Encryption (FHE). This enables privacypreserving data analysis while ensuring the security and confidentiality of patient information. By integrating these independent functions, our system significantly enhances the security and efficiency of health monitoring. It supports critical tasks such as disease detection, patient monitoring, and preliminary intervention, all while upholding stringent data privacy standards. We provided detailed discussions on the system's architecture, hardware configuration, software implementation, and clinical performance. The results highlight the potential of this system to improve patient care through secure and efficient ECG monitoring and analysis. This work represents a significant leap forward in medical technology. By incorporating FHE into both data transmission and storage processes, we ensure continuous encryption of data throughout its lifecycle while enabling real-time disease diagnosis. Our entire architecture is available as open-source, encouraging further research and development in this vital field.

Once analysed, location trajectories can provide valuable insights beneficial to various applications, including urban planning, market analysis, and public health surveillance. However, such data is also highly sensitive, rendering them susceptible to privacy risks in the event of mismanagement, for example, revealing an individual's identity, home address, or political affiliations. Hence, ensuring that privacy is preserved for this data is a priority. One commonly taken measure to mitigate this concern is aggregation. Previous work by Xu et al. in [Trajectory Recovery From Ash: User Privacy Is NOT Preserved in Aggregated Mobility Data (2017)] shows that trajectories are still recoverable from anonymised and aggregated datasets. However, the study lacks implementation details, obfuscating the mechanisms of the attack. Additionally, the attack was evaluated on commercial non-public datasets, rendering the results and subsequent claims unverifiable. This study reimplements the trajectory recovery attack from scratch and evaluates it on two open-source datasets, detailing the preprocessing steps and implementation. Results confirm that privacy leakage still exists despite common anonymisation and aggregation methods but also indicate that the initial accuracy claims may have been overly ambitious. We release all code as open-source to ensure the results are entirely reproducible and, therefore, verifiable. Moreover, we propose a stronger attack by designing a series of enhancements to the baseline attack. These enhancements yield higher accuracies by up to 16%, providing an improved benchmark for future research in trajectory recovery methods. Our improvements also enable online execution of the attack, allowing partial attacks on larger datasets previously considered unprocessable, thereby furthering the extent of privacy leakage. The findings emphasise the importance of using strong privacy-preserving mechanisms when releasing aggregated mobility data and not solely relying on aggregation as a means of anonymisation.

This paper shows the objective necessity of improving the security information systems (SIS) under the development of information and telecommunication technologies. In contrast to existing multiplicative knapsack security information systems based on standard (binary) multiplicative knapsacks KMS, this work, first of all, is about generalization of classic multiplicative knapsack problem to KMG, and, secondly, in the work, mathematical models MMG of security information systems, based on the generalized multiplicative knapsacks AMG, are offered. Necessary and sufficient conditions at which generalized multiplicative knapsack vector is injective over Zp, p≥2 is established.

This paper introduces a novel, multi-dimensional approach to address the evolving challenges in cybersecurity incident response. Our proposed framework uniquely integrates adaptive metrics with a layered security model, providing organisations with a dynamic and context-sensitive tool for building robust response capabilities. We present an innovative integration of proactive threat hunting, real-time threat intelligence, and AI/ML analysis within a cohesive, adaptable structure-a combination not previously explored in incident response literature. This approach not only serves as a comprehensive baseline for managing and responding to incidents but also offers a comparative measure for organisations to continuously evaluate and enhance their cybersecurity postures. Through practical implementation scenarios and future prospects analysis, we demonstrate the framework's unique ability to adapt to the rapidly changing digital landscape, addressing critical gaps in current incident response strategies.

To detect malicious URLs more timely, machine learning based malicious URL detection methods have replaced traditional blacklist methods. These studies aim to improve the accuracy and speed of detection from various aspects such as URL segmentation, URL embedding methods, machine learning models, etc. However, the security issues inherent in these machine learning based malicious URL detection methods have been overlooked. Adversarial example attacks are one of the security issues faced by machine learning based malicious URL detectors. In this paper, we proposed a new adversarial example attack method against malicious URL detection based on machine learning and it has better performance than existing methods. Besides, we compared the robustness of different URL embedding methods and machine learning models with our attack methods and existing attack methods. At last, we analyzed the reasons why our proposed method performs better and the reasons why the context-considered embedding method has high resistance to adversarial example attacks.

Chronic kidney disease is a noticeable health condition that can persist throughout an individual's life, resulting from either kidney malignancy or diminished kidney function. In our study, we delve into the potential of various machine learning methods to offer an early diagnosis of CKD. While previous research has extensively explored this area, our aim is to refine our approach by employing predictive modeling techniques. Initially, we considered 25 variables alongside the class property. The data set used in this study underwent extensive processing, including changing the names of colour for clarity, converting identified colour to numbers, treating unique values with letters handling of partitioned values, fixing incorrect values, filling null values with mean, and encoding categorical values into mathematical notation. In addition, Principal component analysis (PCA) was also employed to lower dimensionality. Based on our findings, the XG Boost classifier performed the best, with an accuracy of 0.993.

The integration of Large Language Models (LLMs) in electricity load forecasting and the enhancement of smart grid accuracy presents a promising avenue for advancing energy management systems. Smart Grids [SGs] are modernized networks designed to deliver efficient, reliable, and economical energy to consumers, necessitating precise electricity consumption forecasting to optimize performance. In this study, we investigate the ability of LLMs to capture complex relationships and patterns in energy consumption data and provide a detailed exploration of their usefulness in electrical demand prediction. Typically, open-source forecasting models suffer from variability in electricity consumption patterns due to external factors such as temperature and market price. Some recent research has shown that LLM models can adapt well to time-series data and can help improve forecasting accuracy. Using prompting and autoregressive generating techniques, LLMs can scale to handle massive data volumes, support bespoke forecasting proposals, and enable smart decision-making in energy systems. This research trend would lead smart gird to be effectively scalable and adapt which shows the role of LLMs in electricity load forecasting frameworks, especially when dealing with EVs charging in a dynamic energy environment.

This study explores advancements in AI-generated image detection, emphasizing the increasing realism of images, including deepfakes, and the need for effective detection methods. Traditional Convolutional Neural Networks (CNNs) have shown success but face limitations in generalization and accuracy, particularly with newer technologies like Diffusion Models. With the evolution of AI image generation models, from CNNs to Generative Adversarial Networks (GANs) and Diffusion Models, detecting synthetic images has become more challenging. Issues include dataset diversity, adversarial attacks, and inconsistencies in pre-processing methods. While state-of-the-art models like CNNs, Vision Transformers (ViTs), and hybrid approaches exist, their accuracy in detecting increasingly sophisticated fake images remains suboptimal. This research proposes a novel hybrid detection model combining CNNs and ViTs with an additional attention mechanism layer. This structure aims to improve the interaction between local and global features, enhancing detection accuracy. The model was trained using the CIFAKE dataset, which contains 120,000 real and AI-generated images. The added attention mechanism enhances feature extraction, addressing limitations in existing models when faced with next generation synthetic images. The hybrid CNN/ViT+Attention model demonstrated improved detection accuracy, achieving 99.77%, surpassing previous methods. This research lays a foundation for stronger AI-generated image detection, helping to mitigate the risks of synthetic image fraud.

Phishing attacks are a significant threat to cybersecurity, particularly among university students who are frequent targets due to their extensive online activities and limited cybersecurity awareness. This study explores the impact of various factors, including threat susceptibility, phishing avoidance behavior, and the use of anti-phishing tools, on students' awareness of phishing attacks. Using a quantitative approach, data were collected from 715 university students worldwide through a structured questionnaire. The findings reveal that while students exhibit a moderate level of awareness about phishing attacks, their reliance on anti-phishing tools remains insufficient. The study identifies a significant positive relationship between the use of anti-phishing tools and increased phishing awareness and avoidance behaviors. Additionally, the research highlights the mediating role of anti-phishing tools in enhancing students' cybersecurity awareness. The results underscore the importance of integrating educational programs and advanced anti-phishing tools to improve students' resilience against phishing attacks. Recommendations for enhancing cybersecurity education and practices among university students are also provided.

Verifiable Timed Signatures (VTS) are cryptographic constructs that enable obtaining a signature at a specific time in the future and provide evidence that the signature is legitimate. This framework particularly finds utility in applications such as payment channel networks, multiparty signing operations, or multiparty computation, especially within blockchain architectures. Currently, VTS schemes are based on signature algorithms such as BLS signature, Schnorr signature, and ECDSA. These signature algorithms are considered insecure against quantum attacks due to the effect of Shor's Algorithm on the discrete logarithm problem. We present a new VTS scheme called VT-Dilithium based on CRYSTALS-Dilithium Digital Signature Algorithm that has been selected as NIST's quantum-resistant digital signature standard and is considered secure against both classical and quantum attacks. Integrating Dilithium into the VTS scheme is a more challenging problem due to its complex mathematical operations (i.e. polynomial multiplications, rounding operations) and large module parameters such as polynomials, polynomial vectors, and matrices. This work aims to provide a comprehensive exposition of VT-Dilithium scheme.

Phishing attacks are among the most prevalent cyber threats, often leading to financial losses, reputational damage, and personal identity crises. This study investigates university students' behavior towards phishing attacks, focusing on the relationship between phishing Awareness and phishing avoidance behavior. Additionally, it examines the moderating effects of gender, age, and qualification on the use of anti-phishing tools, phishing avoidance behavior, and phishing Awareness. Data were collected from 715 university students through a structured questionnaire employing a quantitative approach. The results revealed a strong positive relationship between students' Awareness of phishing attacks and their phishing avoidance behavior. The mediation analysis showed that phishing awareness significantly mediates the relationship between using anti-phishing tools and phishing avoidance behavior. Furthermore, significant differences were observed in phishing Awareness, avoidance behavior, and the use of anti-phishing tools based on gender and age groups. These findings highlight the importance of tailored cybersecurity education programs considering these demographic factors to enhance students' resilience against phishing attacks. It is recommended to develop targeted cybersecurity education programs that focus on increasing phishing awareness and promoting the use of anti-phishing tools among university students, with particular emphasis on addressing demographic differences. The study's findings suggest that improving phishing awareness and tailoring interventions based on gender, age, and educational background can significantly enhance students' ability to avoid phishing attacks, strengthening cybersecurity resilience.

Advanced Metering Infrastructure (AMI) is made up of networks and systems for data collection, analysis, storage, measurement, and utilization related to energy use. Since AMI is a technology that makes the Smart Grid (SG) possible, it is an ideal target for security breaches that might seriously harm infrastructures and privacy. Because of this, creating an effective Key Management System (KMS) is one of the first crucial tasks in developing an AMI, and security has been recognized as one of the most difficult subjects. In order to protect data transfers in an Advanced Metering Infrastructure, we present a scalable and effective key management scheme. This key management system, which is based on an effective Multi-group Key graph technique, may enable broadcast, multicast, and unicast communications. The results of a security and performance analysis, as well as a comparison between our scheme and other recently suggested schemes, demonstrate that our KMS reduces storage overhead relative to current systems while maintaining communication overhead and computational overhead

In recent years, the rapid improvement of deep learning technologies, particularly Generative Adversarial Networks, has led to the proliferation of high-quality synthetic facial images and videos, commonly known as deepfakes. This study aims to evaluate and compare the performance of three prominent deep learning models - ResNet, EfficientNet, and Xception - in detecting synthetic faces. Using the Deepfake Detection Challenge and FaceForensics++ datasets, we systematically assess each model's capability to handle diverse and challenging scenarios, including blurred and dark images. Data augmentation techniques, such as random blurring, brightness adjustment, and contrast enhancement, were employed to improve the models' robustness. Additionally, we applied model-specific optimizations, including the integration of Squeeze-and-Excitation blocks in ResNet, compound scaling in EfficientNet, and multi-scale feature fusion in Xception. These enhancements significantly improved the models' accuracy and resilience against low-quality synthetic data. Our results indicate that EfficientNet and Xception outperform ResNet in both general and adverse conditions, with EfficientNet excelling in high-resolution image processing and Xception showing superior performance in fine-grained feature extraction. Furthermore, the introduction of pre-trained weights, multitask learning frameworks, and dynamic learning rate adjustments during training contributed to the models' enhanced performance.

In recent years, Generative Adversarial Networks (GANs) have emerged as a groundbreaking technique inside the realm of system getting to know, in particular for the technology of synthetic information. Their capacity to research complicated facts distributions and generate high-constancy samples has fueled huge research and innovation in diverse domain names. This summary delves deeper into the mechanisms of GANs, elucidating how the interaction between the generator and discriminator networks results in the introduction of synthetic images that closely resemble real statistics. GANs continuously enhance the greatness of generated samples, pushing the boundaries of what is viable in the artificial facts era. Furthermore, this work explores the potential of visualization data in a variety of industries. Our model achieved more than 90% accuracy in generating the images . The model has wide range of application with a lot of future scope and we are still finding new things every day. It is invented by Ian Goodfellow in 2014 it is making its place in our day to day life. In healthcare, artificial intelligence can facilitate training on medical imaging models without compromising patient privacy. It can also help with risk assessment and fraud detection in the financial sector. It can improve monitoring systems and forensic investigations in security. Moreover, the generation of images has important applications in creative fields such as design and entertainment. This includes creating anime-like visuals, creating realistic faces, and creating unique artwork. By harnessing the potential of GANs, researchers and practitioners can unlock new avenues for innovation and problem solving in a wide range of domains.

This work presents a highly effective strategy for attacking image captioning models through the use of prompt engineering. The objective of this approach is to deliberately guiding the output of LLMs and introduce dynamic noise into the original clean image captions, causing them to be categorized as a different class. Consequently, when the image captioning model is fine-tuned using adversarial captions, it will deteriorate and produce inaccurate captions for clean photos. The novelty of this attack is that it does not require the attacker to perform any model training and only require to prompt the LLMs to generate only a small amount of captions for the attack to be effective. Comprehensive experiments using GPT-3.5 indicate that with only 100 captions created by LLMs with malicious intent can significantly worsen picture captioning model performance by up to over 50% in BLEU metric and over 25% in ROUGE-L and METEOR scores.

Ultra-Reliable Low Latency Communications (URLLC) and OpenRAN are transforming wireless telecommunications. However, concerns regarding security pose significant challenges to their widespread adoption. This paper conducts a series of systematic experiments to uncover hidden vulnerabilities such as performance degradation, potential impact on Quality of Service (QoS) or Quality of Experience (QoE). The experiments are designed to carry out strategic attacks on a Handover AI xApp using four different methods, each with its own unique attack strategy. Our research demonstrates that current policies are inadequate, as URLLC xApps are vulnerable to various types of attacks. This highlights the potential for malicious actors to carry out strategic attacks.

Personal data stores (PDSs) are gaining attention as an alternative to the current systems used by tech giants to manage users' data. PDSs centrally manage user logs and data across services and enable data to be passed on to service providers in accordance with the user's own intentions. Using cloud servers to manage personal information is a natural choice from the perspectives of convenience and cost. Therefore, security is highly important in the PDS model. However, to the best of our knowledge, not enough research on encrypting PDS data has been conducted. This is because merely using conventional encryption schemes could compromise the functionality of PDSs. In this study, we developed a key-aggregate identity-based proxy re-encryption (KA-IB-PRE) scheme combining a key-aggregate cryptosystem technique with an identity-based proxy re-encryption scheme. We demonstrated the syntax and specific configuration of KA-IB-PRE and conducted security proofs. Our approach is highly compatible with PDSs, and a system using KA-IB-PRE called EncPDS can construct a privacy-preserving model that meets their requirements. In particular, it offers advantages in privacy protection compared with existing methods, allowing for selective re-encryption of any data through key aggregation while keeping information about which data to re-encrypt hidden from the cloud.

Most of the devastating cyber-attacks are caused by insiders with access privileges inside an organization. The main reason of insider attacks being more effective is that they don't have many security barriers before they get into the critical resources of the system. Different machine learning techniques have been previously utilized to identify insider threats within cybersecurity domain whereas research done in predicting insider attacks is not significant. Moreover, machine learning models used for prediction and detection face a critical limitation as they require training on labeled datasets, rendering them less effective for real-time data streams which lack threat presence indicators. This work presents an unsupervised machine learning approach that predicts insider threat using behavior analysis for real-time threat data. Patterns are identified in user behavior, to make predictions about benign and malicious insiders. Features are selected by analyzing activities performed. Selected features are utilized to feed machine learning model which extracts anomalous behavior among users, using anomalies in their activity patterns followed by learning methods for threat detection. A dataset that contains selected features from CERT r4.2 is used to make predictions. The performance of Isolation Forest (iForest) is compared with other algorithms of the same category including One-class SVM, Local Outlier Factor (LOF) and DBSCAN to evaluate the new approach. The iForest shows the best performance accuracy 80 percent and recall 84.2 percent.

While deep learning-based recommendation systems have achieved great success, recommendation system models are also at serious risk of intellectual property infringement. Current model watermarking research faces significant challenges in terms of fidelity, invisibility, and efficiency. Additionally, existing model watermarking techniques are predominantly applied to image data, with limited applicability to tabular data. In this paper, we introduce an innovative watermarking framework designed to safeguard the ownership of recommendation system models. Specifically, we verify recommendation system model ownership by embedding a type of backdoor watermark into the training dataset, which does not affect model performance. We have conducted experiments on several classical datasets to validate the reliability and effectiveness of our approach.

The widespread use and the immense user base make Windows systems a prime target for attackers seeking to exploit vulnerabilities and maximize impact. Modern obfuscation techniques enable malware to evade detection tools, allowing it to intrude on systems and carry out malicious activities. Thus, to prevent potential harm to the victim's system, it is crucial to detect malware at the early execution stage and initiate adequate action. However, there is often a trade-off between accuracy and earliness in malware detection, as detecting threats at earlier stages may sometimes come at the cost of reduced detection accuracy. We introduce an early malware detection framework to balance trade accuracy and earliness. Our proposed framework iteratively constructs API call prefix subsequences and applies security-sensitive embedding. The causal sequence encoder transforms these sequences into contextual vectors, which are then classified by a multilayer perceptron. The proposed framework outperforms the state-of-the-art early detection method CTIMD and the early detection method EarlyMalDetect. The proposed framework is able to detect the malicious activity on or before executing 3% of the actual malware sequence.

In many machine learning scenarios, training occurs outside the control of the model sponsor or the entity using the model. A growing concern in such settings revolves around model poisoning and data poisoning-how training is conducted and which data contributes to the process. This paper introduces a protective scheme against model and data poisoning attacks. Leveraging cryptographic primitives such as hashes, signature schemes, and zero-knowledge proofs, the scheme ensures the integrity of the training process. Hashing maintains the continuity of data from authenticated sensors, while signatures validate the data. In the end, zero-knowledge proofs verify the correct model computation by the entity carrying out the training process. By adopting this approach, model sponsors can securely delegate training tasks, guaranteeing the authenticity of the results. Implementation and testing demonstrate the scheme's feasibility, effectively countering data and model poisoning threats.

This paper surveys the potential of contextualized AI in enhancing cyber defense capabilities, revealing significant research growth from 2015 to 2024. We identify a focus on robustness, reliability, and integration methods, while noting gaps in organizational trust and governance frameworks. Our study employs two LLM-assisted literature survey methodologies: (A) ChatGPT 4 for exploration, and (B) Gemma 2:9b for filtering with Claude 3.5 Sonnet for full-text analysis. We discuss the effectiveness and challenges of using LLMs in academic research, providing insights for future researchers.

Integrating Groundbreaking advancements in AI, like language models, interpretable AI, and machine learning, opens up a world of exciting new possibilities. Machine Learning algorithms hold the potential to enhance threat detection & response. The Evolving face of cybersecurity and Modern cyber threats are complex and well crafted; hence, conventional cybersecurity mechanisms show difficulty in staying relevant. LLMs, especially based on Transformer architecture will noticeably increase the accuracy and speed of detecting threats. Transparency and trust are increased by XAI approaches like SHAP and LIME, which offer facts about ML model predictions. This paper explores the literature that demonstrates the integration between XAI and LLMs in cybersecurity, exemplifying how this trinity of models has the potential to help attenuate errors producing Reduced false positives and improve how we detect threats. Thinking about the possibilities the challenges including performance Explainability trade-offs, the need for common evaluation metrics, and the black-box nature of AI Models, remain in place. Solving these will help to enhance AI-driven solutions in cybersecurity.

Clean water is a basic human need and necessary for environmental sustainability, particularly in rural areas with limited supplies of water where basic methods for monitoring water quality are seriously impeded. Combining explainable artificial intelligence (XAI) and large language models (LLMs) with remote sensing offers a possible solution. In this review, we examine an innovative framework that leverages satellite and UAV data to assess water quality through pioneering AI techniques. The proposed framework strongly emphasizes using LLMs for better data processing and Perspective formation, with XAI techniques to ensure trust at both stakeholder level and model visibility. Application of this approach could significantly improve real-time monitoring and notably valuable outcomes to conserve water supplies in resource-constrained environments. The framework is designed to be scalable and flexible to fit the specific environmental context of either rural environments or socioeconomic variables to ensure any water quality-monitoring program remains sustainable over time. This analysis shows how we can combine such emergent technology as AI with remote sensing to tackle large-scale water quality problems and open up new ways of monitoring our precious aquatic resources.

Active Directory is adopted by many organizations today. However, not many know how to securely implement it. This is concerning, particularly with the increasing prevalence of ransomware attacks. A ransomware attack inevitably affects organizations' business operations, data loss, and causes financial loss. Our approach aims to enable analysis of technical routes and behavioral aspects of attackers' "pre-ransomware" actions in an AD network, while taking their capabilities and attack surfaces into account for realistic scenarios. By integrating simulations of vulnerable AD networks with an SIEM server, we demonstrate how AD attacks are carried out, with the goal of collecting raw data on such attacks for future study for improvement of organizational detection capabilities.

Protecting information flow, data and assets is paramount to every establishment. Therefore, enterprise security architecture design is essential in achieving this protection as it directly implements enterprise security policies. Existing research revealed that researchers have made little effort to investigate inference security challenges to enterprise security architecture design and to assess how the existing security architecture models fare against inference attacks. It was also discovered that existing security architecture models are too old and susceptible to inference attacks. Hence, this research explores a novel solution for designing effective enterprise security architecture and addressing inference attacks.

Quantum-safe public key exchange protocols face significant challenges, both in hardware-based and software-based approaches. Quantum key distribution, which relies on specialized quantum hardware, presents a significant barrier to widespread adoption due to its high cost and limited scalability. Conversely, software-based solutions using post-quantum algorithms introduce their complications, such as increased resource demands and larger ciphertexts. Furthermore, the security of these post-quantum algorithms remains relatively untested, which has led to the emerging trend of hybrid deployment, combining classical and quantum-resistant techniques to hedge against potential vulnerabilities.

In this work, we address these problems by proposing a novel quantum-safe symmetric key agreement (SKA) protocol that is both lightweight and scalable. Our approach involves a hybrid mechanism, leveraging secret strings distributed through a combination of classical and quantum public key pairs during the initial key exchange. This hybrid approach enhances security by utilizing both quantum-resistant algorithms and classical methods, mitigating the risks associated with the nascent nature of post-quantum cryptography. After the initial key exchange, the protocol completes the process using a quantum-safe AES symmetric key, ensuring both security and efficiency. All communications are securely authenticated over classical TLS, making our solution compatible with existing infrastructure.

The contributions of this work are threefold. \textit{First}, we demonstrate that our protocol incurs minimal performance overhead, with only 99ms for purely quantum SKA and 199ms for the hybrid version, compared to classical SKA protocol. \textit{Second}, our SKA protocol remains robust under various network conditions, including delays, packet losses, and bandwidth variations, maintaining small and consistent overheads. \textit{Third}, we show that our solution is highly scalable, with an overhead of only one second for every additional five concurrent users, and that performance improves significantly with increased computational resources-achieving a 50-60% improvement when scaling from two to four CPUs. Additionally, our security evaluations confirm that the protocol provides consistent and sufficient randomness throughout the key agreement process, ensuring quantum-resistance at every stage.

Biometric authentication systems have become an inseparable part of society. This popularity is owing to the fact that biometric traits are immutable. However, applications using these sensitive biometric traits must treat this crucial information carefully. Otherwise, their leakage can threaten the security and privacy of an individual. This paper proposes a privacy-preserving biometric authentication system based on iris data. In the proposed framework, the homomorphic properties are exploited to process data while it is encrypted. There is no leakage of sensitive data throughout the entire process, even when utilizing the services of third-party cloud service providers (CSPs). Authentication is carried out fully within the encrypted domain. Experiments have been conducted to validate the robustness of the proposed framework. Additionally, the time complexity of the proposed framework is minimized compared to other state-of-the-art approaches.

The prominence of the Internet of Things (IoT) has surged in recent years due to technological advancements, increased connectivity, and the widespread adoption of smart devices. Currently, IoT solutions are pervasive across various sectors, including smart homes, wearable devices, healthcare, transportation, and industrial automation. The Long Range Wide Area Network (LoRaWAN) is a wireless communication protocol designed to provide reliable connection over long distances, spanning many kilometers, using low-power and wide-area networks (LPWANs) often used in IoT applications. However, LoRaWAN is vulnerable to bit-flipping attacks, where transmitted data is intercepted and maliciously altered by flipping specific bits in the message payload. Such attacks have the potential to undermine the trustworthiness and dependability of the system. In this paper, we propose a novel security enhancement for LoRaWAN. Our scheme incorporates a hash digest along with the end device's payload, enabling the application server to detect any malicious alterations in the received content using the hash value. We analyze the security of the proposed scheme against bit-flipping attacks and measure its performance through experiments in a real testbed. Our observations indicate that the proposed mechanism not only secures communication against these attacks but also incurs minimal overhead in terms of power consumption, transmission overhead, and latency.

In the rapidly evolving realm of electric vehicle technology, safeguarding the cybersecurity of both electric vehicles and their charging infrastructure has become fundamental. The integration of electric vehicles and their charging stations into the broader grid introduces complex cybersecurity challenges, necessitating robust incident response strategies. Traditional cybersecurity playbooks often fall short in addressing the unique vulnerabilities associated with electric vehicles and their charging systems. The lack of publicly available community playbooks tailored to these needs leaves the electric vehicle ecosystem vulnerable to cyber threats that could compromise user privacy, vehicle functionality, and grid stability. In response to this, the project undertakes the creation of a foundational playbook for electric vehicle and electric vehicle charging station incident response, addressing a significant void in current cybersecurity practices. This paper introduces a Playbook Manager and Visualizer application, called EV-IRP, designed to enable users to upload, manage, and visualize electric vehicle and electric vehicle charging station incident response playbooks efficiently. Utilizing Python, Tkinter for GUI development, SQLite for database management, and Graphviz for visualization, the application facilitates a dynamic and responsive approach to maintaining upto- date incident response strategies. The application is expected to streamline the management of incident response playbooks through procedure visualization, enhancing the cybersecurity posture of electric vehicle infrastructure. By converting textual playbook procedures into easily understandable diagrams, it facilitates a clearer understanding of response steps among users, thereby enhancing the overall efficiency and efficacy of incident response practices. Additionally, the research and development process, informed by a comprehensive literature review, contributes to the academic and practical understanding of cybersecurity best practices for electric vehicle technologies.

Due to massive growth in IoT devices in recent years, security of these devices is a major concern. IoT applications span across a number of fields including but not limited to smart cities, intelligent agriculture systems, and the innovative industry. Despite its benefits, cybersecurity challenges have increased significantly in IoT environments. The lack of resource capacity and sophisticated security measurement exposes IoT devices to large number of recent attacks. A strong intrusion detection system (IDS) is the best way to secure IoT devices. Various studies have shown that the current IDS fails to detect modern malware in IoT environments. Some datasets do not have complex scenarios of attack. There are limitations of current datasets, or heterogeneous data of the IoT environment, such as KDD99, NLS_KDD, and UNSW_NB15. In addition, these datasets do not include an operating system and network monitoring audits. This paper comprehensively compares five machine-learning models on the recent EDGE-IIoT dataset. We examine these machine learning methods on Binary and Multiclass class IDS and the security challenges in managing current and future attacks in an IoT environment.

With the widespread of smart devices, a massive number of messages are sent through message applications every day. Currently, most message applications have been designed for single-receiver and group-receiver scenarios. Moreover, the sender device may send messages to multiple-receiver devices selected from the group device set. In this paper, we propose a lightweight hybrid signcryption scheme supporting the single-receiver and multiple-receiver scenarios, which are compatible with the group-receiver scenario. The formal verification results show that our scheme can offer confidentiality and authentication and is resistant to key-compromise impersonation (KCI) attacks. The evaluation results show that it provides an efficient solution with lower compute time and smaller payload size, which is suitable for lightweight message delivery.

The global outbreak of new coronavirus infections has increased the demand for smartphone-based contact tracing. In particular, much research has been conducted on contact tracing using blockchain, which is decentralized and hard to tamper with data. However, contact tracing requires the centralized collection of information for tracing users, and this essence remains the same even when using blockchain. In this study, we propose a blockchain-based contact tracing method that mitigates the centralized information collection and considers user privacy so that the administrator cannot trace the information. In this scheme, only the application knows the wallet address corresponding to the user, so even the trust authority cannot determine the user's identity from the wallet address. This method is capable of notifying users who may be witnesses at a crime scene by using smart contracts. Users can also prove their innocence if suspected of a crime. Furthermore, we will demonstrate the effectiveness of this method by conducting a demonstration experiment using an Ethereum testnet and two smartphones.

The rapid growth and widespread application of Internet of Things (IoT) devices have resulted in a significant increase in the demand for data transmission. However, IoT devices often operate in unstable environments and are constrained by their limited computing and communication capabilities, which makes stable and reliable data transmission a major challenge. Due to the diversity and complexity of IoT platforms, existing systems struggle to provide flexible and scalable solutions. To address these issues, we propose a system based on blockchain and smart contracts that enhances data transmission redundancy by integrating external collaborative devices, thereby ensuring both transmission stability and security. The proposed system achieves efficient identity verification through the use of certificates and smart contracts, while also promoting the participation of temporary devices through incentive mechanisms. Compared to existing methods, our system not only enhances security but also optimizes resource utilization and reduces operational costs.

The modern food supply chain often involves multiple layers of participants spread across different countries and continents. This complex system offers significant benefits to businesses worldwide; however, it also presents several challenges. One major problem is the inability to trace the product flow back to its origin, a critical issue in many industries. Another issue is the lack of trust among supply chain participants. Blockchain technology can help address these and other challenges faced by the supply chain industry. However, it is surprising that, globally, there are still not many examples of the technology's adoption, with most projects remaining in the pilot stage. This paper explores the field of custom blockchain design tailored to specific applications, with a focus on supply chain operations in the food industry. It includes the development of a lightweight yet fully featured Python prototype for a decentralized blockchain system. In addition to common features like block validation and state updates, the prototype includes a newly designed type of transaction tailored specifically for supply chain operations. These transactions eliminate the need for smart contracts, making the system more lightweight compared to general-purpose blockchain platforms such as Ethereum and less prone to security vulnerabilities. The prototype is designed as a public blockchain network, with Proof of Work selected as the consensus algorithm. The novelty of this research work lies in advancing the concept of a custom blockchain solution for the food industry. The key elements of the prototype have been unit tested. The overall evaluation was completed using a Python script that simulates product flow through an example supply chain, allowing product provenance to be determined by tracing the product flow back to its origin.

This The rise of autonomous vehicles (AVs) brings with it the need for secure and privacy-preserving machine learning models. Federated learning (FL) allows AVs to collaboratively train models while keeping raw data localized. However, traditional FL systems are vulnerable to security threats, including adversarial attacks, data breaches, and dependency on a central aggregator, which can be a single point of failure. To address these concerns, this paper introduces a peer-to-peer decentralized federated learning system that integrates lightweight blockchain technology and Binius Zero-Knowledge Proofs (ZKPs) to enhance security and privacy. In this system, Binius ZKPs ensure that model updates are cryptographically verified without exposing sensitive information, guaranteeing data confidentiality and integrity during the learning process. The lightweight blockchain framework secures the network by creating an immutable, decentralized record of all model updates, thus preventing tampering, fraud, or unauthorized modifications. This decentralized approach eliminates the need for a central aggregator, significantly enhancing system resilience to attacks and making it suitable for dynamic environments like AV networks. Additionally, the system's design includes Byzantine resilience, providing protection against adversarial nodes and ensuring that the global model aggregation process remains robust even in the presence of malicious actors. Extensive performance evaluations demonstrate that the system achieves low-latency, scalability, and efficient resource usage while maintaining strong security and privacy guarantees, making it an ideal solution for real-time federated learning in autonomous vehicle networks. The proposed framework not only ensures privacy but also fosters trust among participants in a fully decentralized environment.

Cybersecurity education incorporates a variety of teaching methods such as traditional lectures, lectures combined with hands-on exercises, and concept maps. One of the most well- known instructional methods is the use of lectures supplemented by hands-on activities. However, often these exercises either lack a strong connect with the lecture material or invariably lead students step-by-step in predetermined tasks, thereby hindering critical thinking and problem-solving skills. Hence, the instructional method falls short on providing students with a comprehensive understanding of complex and often associated cybersecurity concepts as encountered in real-world security incidents. The authors propose that a problem-based learning (PBL) approach can effectively address these gaps and improve cybersecurity education learning outcomes. This paper presents an application of PBL approach for teaching cybersecurity and ethical hacking in community colleges that play a crucial role in meeting the demand for cybersecurity professionals, but often face several challenges to effectively introduce cybersecurity concepts in their curriculum. Through this research, an existing course on ethical hacking is redesigned using the PBL pedagogy and offered to community college students. The course involves several PBL modules that are developed to cover all key aspects of ethical hacking and implemented using open-source software's. Each PBL module is based on a real-world cybersecurity incident and mapped to the MITRE ATT&CK framework. An external independent evaluation is conducted to assess the effectiveness of the proposed teaching methodology. Overall, the obtained results positively impact students' critical thinking, problem-solving, and communication skills, along with facilitating their understanding of key cybersecurity concepts. 100% of students reported that they enjoyed the PBL exercises. 75% of the students believed that PBL enhanced their learning of key concepts to a great extent, and remaining 25% believed that their learning of key concepts was somewhat enhanced.

Cyber security is becoming more complex due to the exponential growth of interconnected systems and the global threat landscape. To mitigate those risks, there is a need for a skilled cyber security workforce that can navigate the complex decision-making in rapidly evolving cyberspace. Artificial intelligence (AI) is very fast adopted into cyber defence operations. However, we can not effectively train human-AI assisted cyber defence operators, without understanding the underlying learning theory and eco-systems. Cyber security exercises (CSXs) are popular teaching methods for cyber-readiness; however, applying learning analytics (LA) methods and AI-based approaches in an exercise design and implementation is still in the early stages. We propose a holistic Human-AI interaction model within the LA and CSX context. The model brings together elements and processes of human-AI interactions, but also cyber ranges, cyber security and LA tools and a wider lense of multimodal learning analytics, exercise life-cycle and overall pedagogical approach taken. We also discuss the opportunities and challenges for LA and AI in the context of cyber security training. We analyse the role of AI from the learning, instruction, and administration lens in cyber security training, specifically in the exercises. We aim to stimulate further discussions on the future of human-AI collaboration and how to enhance cyber security training with novel LA and AI capabilities.

This paper presents the development of a security scoring platform, with the aim of enhancing the organization's defenses against ransomware attacks. The platform features a security score calculator that leverages on a customizable security scoring algorithm that evaluates the adequacy of the organization's security practices and measures, whilst also providing remediation measures for areas of improvement to allow security teams to quickly respond to threats. Users of the platform will also be able to manage assets, past historical scores and export data for comprehensive security assessments. This dual approach of real-time security scoring and predictive anomaly detection could fortify the organization's security posture in the evolving cyber threat landscape and allow for proactive, responsive measures to mitigate risks effectively.

This paper aims to analyze and provide insights on vulnerabilities being exploited by ransomware attacks. This paper offers users a comprehensive overview of exploited Common Vulnerabilities and Exposures (CVE). Our analysis includes displaying charts on various aspects such as the top ransomware gangs, the most frequently abused CVE IDs, comparative information before and after the year 2021, common attack vectors, and more. Key findings from our research show that the top ransomware gangs include Conti and Cerber. The most exploited CVE IDs are CVE-2021-34473 and CVE-2021-34523, which correspond to the Microsoft Exchange Server Remote Code Execution Vulnerability and the Microsoft Exchange Server Elevation of Privilege Vulnerability, respectively. Additionally, we identified common attack vectors such as network attacks, application attacks, and data breaches, with instance counts of 10,091, 7,087, and 1,115, respectively. These insights provide valuable information for understanding and mitigating ransomware threats.