Advancements in computing, AI, and connectivity have revolutionized industries, improving efficiency and innovation, but the reliance on sensitive data raises privacy and security concerns. Specifically, energy-intensive sectors such as buildings, a major contributor to global energy use and CO2emissions, require accurate prediction of energy demand for efficiency and sustainability. In this paper, we present federated learning (FL) and edge computing for secure, decentralized energy modeling in multi-building environments, preserving data privacy. A multibuilding energy dataset was developed for FL applications, and an edge-enabled FL framework with a hybrid deep learning (DL) model was proposed for energy prediction. Furthermore, we performed a comparative analysis against recent DL models presented in the literature for energy consumption prediction using FL. Our proposed model got the lowest error scores of 11.74, 9.46, 18.42, and 11.72 for RMSE, MAE, MAPE, and Std Dev, respectively. The results demonstrate the superior performance and efficiency of the proposed model and highlight the potential of the framework to improve energy management, reduce emissions, and support sustainable urban development.

The increasing complexity and connectivity of Internet of Things (IoT) environments have made them targets for advanced cyber-attacks. Recently, Deep Learning (DL)-based intrusion detection systems have shown remarkable success in identifying malicious activities within IoT traffic. In particular, Graph Neural Networks (GNNs) have emerged as effective solutions. However, GNNs come with inherent challenges, including high computational complexity and a black-box nature, which limit their transparency and interoperability. In this paper, we introduce a hybrid framework that combines a GNN model, named AD-MGCN, with a Large Language Model (LLM) to address these limitations. The proposed AD-MGCN leverages a Markov-based multi-step diffusion process to enhance feature propagation, reduce noisy edges, and improve classification performance across both frequent and rare attack types. In addition, a fine-tuned instruction-based LLM (Falcon-7B Instruct) generates natural language explanations that translate model predictions into human-understandable insights. We evaluated our framework on the Edge-IIoTset dataset, which includes diverse IoT attack scenarios. The experimental results show that AD-MGCN achieves an accuracy of 97.38%, significantly outperforming the baseline GCN models. Furthermore, the LLM explanations achieve an average clarity score of 4.2/5 in expert evaluations, improving the transparency of the model for cybersecurity analysts. These results demonstrate the potential of AD-MGCN as a reliable, efficient, and interpretable solution for securing modern IoT ecosystems.

From the first attack in 1989, to date, it is evident that ransomware is highly destructive. Today the vast majority of research on ransomware detection is focused on the use of AI techniques. While the use of these techniques is very effective, they should not be considered an infallible solution for ransomware detection. As with any solution, AI implementations do have shortcomings of their own; compute resource constraints, collation of training data, data poisoning, and data privacy, to name a few. This paper aims to identify whether traditional methods can still effectively detect ransomware in scenarios where AI solutions may not be viable. Typically, there are three main categories of detection; signature-based, behaviour-based, & deception-based. This paper focuses on deception-based detection, using honey files. Three detection solutions have been implemented on two isolated VMs, one running Windows 10, the other Linux Mint. The solutions include RansomwareLocker, for the Linux VM, R-Locker and 4663 Windows event monitoring on the Windows 10 VM. With these solutions implemented, ransomware samples were executed in turn, up to three times, allowing an initial ‘out of the box’ test run and two subsequent tests after necessary configuration changes were made. Overall, from the ransomware samples chosen and detection solutions implemented, deception-based detection proves to be a promising approach. Testing resulted in two of the three solutions ultimately achieving a 100% detection rate. However, throughout the experiment, it is evident that this approach is not a silver bullet, and very dependent on the configuration of the solutions. Therefore, whether AI-based or traditional, a defence-in-depth approach remains best.

Alzheimer's disease is a severe brain disorder that causes harm in various brain areas and leads to memory damage. The limited availability of labeled medical data poses a significant challenge for accurate Alzheimer's disease detection. There is a critical need for effective methods to improve the accuracy of Alzheimer's disease detection, considering the scarcity of labeled data, the complexity of the disease, and the constraints related to data privacy. To address this challenge, our study leverages the power of Big Data in the form of pre-trained Convolutional Neural Networks (CNNs) within the framework of Few-Shot Learning (FSL) and ensemble learning. We propose an ensemble approach based on a Prototypical Network (ProtoNet), a powerful method in FSL, integrating various pre-trained CNNs as encoders. This integration enhances the richness of features extracted from medical images. Our approach also includes a combination of class-aware loss and entropy loss to ensure a more precise classification of Alzheimer's disease progression levels. The effectiveness of our method was evaluated using two datasets, the Kaggle Alzheimer dataset, and the ADNI dataset, achieving an accuracy of 99.72% and 99.86%, respectively. The comparison of our results with relevant state-of-the-art studies demonstrated that our approach achieved superior accuracy and highlighted its validity and potential for real-world applications in early Alzheimer's disease detection.

In the Internet of Medical Things (IoMT), de novo peptide sequencing prediction is one of the most important techniques for the fields of disease prediction, diagnosis, and treatment. Recently, deep-learning-based peptide sequencing prediction has been a new trend. However, most popular deep learning models for peptide sequencing prediction suffer from poor interpretability and poor ability to capture long-range dependencies. To solve these issues, we propose a model named SeqNovo, which has the encoding-decoding structure of sequence to sequence (Seq2Seq), the highly nonlinear properties of multilayer perceptron (MLP), and the ability of the attention mechanism to capture long-range dependencies. SeqNovo use MLP to improve the feature extraction and utilize the attention mechanism to discover key information. A series of experiments have been conducted to show that the SeqNovo is superior to the Seq2Seq benchmark model, DeepNovo. SeqNovo improves both the accuracy and interpretability of the predictions, which will be expected to support more related research.

With the widespread use of the Internet of Things (IoT), securing the storage and transmission of multimedia content across IoT devices is a critical concern. Chaos-based Pseudo-Random Number Generators (PRNGs) play an essential role in enhancing the security of image encryption algorithms. This paper introduces a novel 1-dimensional cosine-modulated-polynomial chaotic map to be used as a PRNG in image encryption algorithms. The proposed map utilizes a cosine function to modulate the outcome of a polynomial expression, resulting in complex chaotic behaviour. The designed map acts as a self-modulating system and offers a larger chaotic range, reduced structural complexity, and enhanced chaotic properties, such as aperiodicity, unpredictability, ergodicity, and sensitivity to control parameters and initial conditions, in comparison to the traditional 1-dimensional chaotic maps. An extensive evaluation is performed to gauge the chaotic behaviour of the proposed map, including bifurcation diagrams, chaotic trajectory analysis, fixed point and stability analysis, Lyapunov Exponent, Kolmogorov Entropy and NIST SP800-22 tests demonstrating its effectiveness to be used as a secure PRNG in image encryption algorithms.

Blockchain (BC) and Information for Operational and Tactical Analysis (IOTA) are distributed ledgers that record a huge number of transactions in multiple places at the same time using decentralized databases. Both BC and IOTA facilitate Internet-of-Things (IoT) by overcoming the issues related to traditional centralized systems, such as privacy, security, resources cost, performance, and transparency. Still, IoT faces the potential challenges of real-time processing, resource management, and storage services. Edge computing (EC) has been introduced to tackle the underlying challenges of IoT by providing real-time processing, resource management, and storage services nearer to IoT devices on the network’s edge. To make EC more efficient and effective, solutions using BC and IOTA have been devoted to this area. However, BC and IOTA came with their pitfalls. This survey outlines the pitfalls of BC and IOTA in EC and provides research directions to be investigated further.

Deep learning-based disease diagnosis applications are essential for accurate diagnosis at various disease stages. However, using personal data exposes traditional centralized learning systems to privacy concerns. On the other hand, by positioning processing resources closer to the device and enabling more effective data analyses, a distributed computing paradigm has the potential to revolutionize disease diagnosis. Scalable architectures for data analytics are also crucial in healthcare, where data analytics results must have low latency and high dependability and reliability. This study proposes a microservices-based approach for IoT data analytics systems to satisfy privacy and performance requirements by arranging entities into fine-grained, loosely connected, and reusable collections. Our approach relies on federated learning, which can increase disease diagnosis accuracy while protecting data privacy. Additionally, we employ transfer learning to obtain more efficient models. Using more than 5800 chest X-ray images for pneumonia detection from a publically available dataset, we ran experiments to assess the effectiveness of our approach. Our experiments reveal that our approach performs better in identifying pneumonia than other cutting-edge technologies, demonstrating our approach's promising potential detection performance.

In the era of Industrial IoT (IIoT) and Industry 4.0, ensuring secure data transmission has become a critical concern. Among other data types, images are widely transmitted and utilized across various IIoT applications, ranging from sensor-generated visual data and real-time remote monitoring to quality control in production lines. The encryption of these images is essential for maintaining operational integrity, data confidentiality, and seamless integration with analytics platforms. This paper addresses these critical concerns by proposing a robust image encryption algorithm tailored for IIoT and Cyber-Physical Systems (CPS). The algorithm combines Rule-30 cellular automata with chaotic scrambling and substitution. The Rule 30 cellular automata serves as an efficient mechanism for generating pseudo-random sequences that enable fast encryption and decryption cycles suitable for realtime sensor data in industrial settings. Most importantly, it induces non-linearity in the encryption algorithm. Furthermore, to increase the chaotic range and keyspace of the algorithm, which is vital for security in distributed industrial networks, a hybrid chaotic map, i.e., logistic-sine map is utilized. Extensive security analysis has been carried out to validate the efficacy of the proposed algorithm. Results indicate that our algorithm achieves close-to-ideal values, with an entropy of 7.99 and a correlation of 0.002. This enhances the algorithm's resilience against potential cyber-attacks in the industrial domain.

The synergy between the Internet of Things (IoT) and big data technologies has resulted in the great development of multiple smart applications in varied fields such as energy management, environmental monitoring, elderly healthcare, etc. Due to the increasing demand for smart applications, opting for a flexible and scalable software architecture that supports and accelerates the development of these applications is in dire need nowadays. As an effective solution to continuously maintain, upgrade, and scale IoT-based applications, the microservices paradigm has been adopted as an architectural style allowing to provide several enhancements in terms of independent deployment, modularity, containerization, loose coupling, etc. These advantages provided by microservices impact also the efficiency of the analytics conducted by the IoT-based systems. In this view, the current research aims to address a survey about the adoption of the microservices paradigm for supporting data analytics in IoT applications. For this purpose, first of all, the theoretical concepts related to IoT, data analytics, and microservices are briefly presented. Second, relevant microservices-based solutions for data analytics in IoT applications are reviewed and discussed. Third, the challenges and opportunities offered by the microservices paradigm in the IoT context are outlined and explained. The present study will pave the way for promising future research on the issues of integrating the emerging microservices technology in data analytics and IoT applications.

Abstract Wireless sensor networks (WSNs) have recently attracted greater attention worldwide due to their practicality in monitoring, communicating, and reporting specific physical phenomena. The data collected by WSNs is often inaccurate as a result of unavoidable environmental factors, which may include noise, signal weakness, or intrusion attacks depending on the specific situation. Sending high‐noise data has negative effects not just on data accuracy and network reliability, but also regarding the decision‐making processes in the base station. Anomaly detection, or outlier detection, is the process of detecting noisy data amidst the contexts thus described. The literature contains relatively few noise detection techniques in the context of WSNs, particularly for outlier‐detection algorithms applying time series analysis, which considers the effective neighbors to ensure a global‐collaborative detection. Hence, the research presented in this article is intended to design and implement a global outlier‐detection approach, which allows us to find and select appropriate neighbors to ensure an adaptive collaborative detection based on time‐series analysis and entropy techniques. The proposed approach applies a random forest algorithm for identifying the best results. To measure the effectiveness and efficiency of the proposed approach, a comprehensive and real scenario provided by the Intel Berkeley Research Laboratory has been simulated. Noisy data have been injected into the collected data randomly. The results obtained from the experiment then conducted experimentation demonstrate that our approach can detect anomalies with up to 99% accuracy.

IoT applications have extended new concepts to smart technologies in different vital sectors such as healthcare, education, agriculture, energy management and control, etc. IoT aims to interconnect several intelligent devices over the Internet to control, store, exchange, and analyze collected data. However, the main problem in IoT environments is that they present numerous potential vulnerabilities that can be the origin of security attacks. In this context, this paper proposes a novel framework, named SUBLμME, which stands for SecUre BLockchain as a service and μservices-based fraMework for IoT Environments. This framework integrates BaaS and microservices technologies that have the predominant characteristic of offering reusable and reconfigurable security features implemented as independent services that can be reused for multiple IoT applications. SUBLμME is implemented and validated by a simulated smart home environment. The results of the simulation experiments showed that SUBLμME brings several security improvements in terms of performance, data integrity, data packages’ validation, access control, and efficient and secure data transmission and storage.

The COVID-19 pandemic is causing a global health crisis. Public spaces need to be safeguarded from the adverse effects of this pandemic. Wearing a facemask becomes one of the effective protection solutions adopted by many governments. Manual real-time monitoring of facemask wearing for a large group of people is becoming a difficult task. The goal of this paper is to use deep learning (DL), which has shown excellent results in many real-life applications, to ensure efficient real-time facemask detection. The proposed approach is based on two steps. An off-line step aiming to create a DL model that is able to detect and locate facemasks and whether they are appropriately worn. An online step that deploys the DL model at edge computing in order to detect masks in real-time. In this study, we propose to use MobileNetV2 to detect facemask in real-time. Several experiments are conducted and show good performances of the proposed approach (99% for training and testing accuracy). In addition, several comparisons with many state-of-the-art models namely ResNet50, DenseNet, and VGG16 show good performance of the MobileNetV2 in terms of training time and accuracy.

Recently, Convolutional Neural Networks (CNNs) have made a great performance for remote sensing image classification. Plant recognition using CNNs is one of the active deep learning research topics due to its added-value in different related fields, especially environmental conservation and natural areas preservation. Automatic recognition of plants in protected areas helps in the surveillance process of these zones and ensures the sustainability of their ecosystems. In this work, we propose an Enhanced Randomly Initialized Convolutional Neural Network (ERI-CNN) for the recognition of columnar cactus, which is an endemic plant that exists in the Tehuacán-Cuicatlán Valley in southeastern Mexico. We used a public dataset created by a group of researchers that consists of more than 20000 remote sensing images. The experimental results confirm the effectiveness of the proposed model compared to other models reported in the literature like InceptionV3 and the modified LeNet-5 CNN. Our ERI-CNN provides 98% of accuracy, 97% of precision, 97% of recall, 97.5% as f1-score, and 0.056 loss.

In recent years, the Internet of Things (IoT) technology has led to the emergence of multiple smart applications in different vital sectors including healthcare, education, agriculture, energy management, etc. IoT aims to interconnect several intelligent devices over the Internet such as sensors, monitoring systems, and smart appliances to control, store, exchange, and analyze collected data. The main issue in IoT environments is that they can present potential vulnerabilities to be illegally accessed by malicious users, which threatens the safety and privacy of gathered data. To face this problem, several recent works have been conducted using microservices-based architecture to minimize the security threats and attacks related to IoT data. By employing microservices, these works offer extensible, reusable, and reconfigurable security features. In this paper, we aim to provide a survey about microservices-based approaches for securing IoT applications. This survey will help practitioners understand ongoing challenges and explore new and promising research opportunities in the IoT security field. To the best of our knowledge, this paper constitutes the first survey that investigates the use of microservices technology for securing IoT applications.

Summary Wireless sensor networks (WSNs) consist of small sensors with limited computational and communication capabilities. Reading data in WSN is not always reliable due to open environmental factors such as noise, weakly received signal strength, and intrusion attacks. The process of detecting highly noisy data is called anomaly or outlier detection. The challenging aspect of noise detection in WSN is related to the limited computational and communication capabilities of sensors. The purpose of this research is to design a local time‐series‐based data noise and anomaly detection approach for WSN. The proposed local outlier detection algorithm (LODA) is a decentralized noise detection algorithm that runs on each sensor node individually with three important features: reduction mechanism that eliminates the noneffective features, determination of the memory size of data histogram to accomplish the effective available memory, and classification for predicting noisy data. An adaptive Bayesian network is used as the classification algorithm for prediction and identification of outliers in each sensor node locally. Results of our approach are compared with four well‐known algorithms using benchmark real‐life datasets, which demonstrate that LODA can achieve higher (up to 89%) accuracy in the prediction of outliers in real sensory data.

Building a composite application based on Web services has become a real challenge regarding the large and diverse service space nowadays. Especially when considering the various functional and non-functional capabilities that Web services may afford and users may require. In this paper, we propose an approach for facilitating Web service selection according to user requirements. These requirements specify the needed functionality and expected QoS, as well as the composability between each pair of services. The originality of our approach is embodied in the use of Relational Concept Analysis (RCA), an extension of Formal Concept Analysis (FCA). Using RCA, we classify services by their calculated QoS levels and composability modes. We use a real case study of 901 services to show how to accomplish an efficient selection of services satisfying a specified set of functional and non-functional requirements.

The new paradigm for distributed computing over the Internet is that of Web services (WSs). One of the key ideas of this new paradigm is the ability to create value-added Service-Based Applications (SBAs) by composing pre-existing services. Building SBAs necessitates the discovery and the selection of the most appropriate WSs that fit closely users' functional and non-functional requirements. Due to the large number of WSs that are advertised over public and private registries and the various functional and non-functional capabilities that are required by users, discovery and selection of WSs have become a real challenge nowadays. In this paper, we present a WS composition approach that is built upon both perspectives: intentional and operational. In the intentional perspective, we propose to model users' requirements for SBAs using the MAP formalism and specify the required WSs using an Intentional Service Model (ISM). In the operational perspective, we propose to discover the required WSs by querying the service search engine Service-Finder and select the most appropriate WSs by using many-valued concept lattices. To validate our approach, we use an analytical technique that is the monitoring to verify that the selected WSs assure the required users' non-functional capabilities.