Abstract The Industrial Internet of Things (IIoT) applications have been recognized as an advancement of the conventional wireless network that concentrates on incorporating processes and machines specifically for industrial applications. These Industrial applications frequently use centralized machine learning (ML) approaches not only to enhance their functionality but also to evaluate sensor data for a variety of purposes, including digitizing operations in manufacturers, forecasting maintenance requirements in industrial equipment, and detecting anomalies for security monitoring, they may adversely affect overall system performance due to high cost of computing power and privacy concerns, as so much data is stored on a cloud server. Federated Learning (FL) has emerged as a new benchmark for centralized ML methods. It sends models to user devices without transferring private data to third-party or central servers; it is one of the promising solutions to data leakage issues. This work introduces a comprehensive overview of the advancements, challenges, and future directions in FL adoption with edge devices. It covers security threats and mitigation strategies, emphasizing its categories, privacy and concerns, communication overhead obstacles, heterogeneity issues, aggregation techniques, and associated development tools. This review paper delves into FL-related topics, including system platforms, offering a comprehensive overview of best practice systems in real-world FL applications. To ensure security in IIoT applications, reviewing threats and mitigation strategies by integrating FL with state-of-the-art technologies such as blockchain, federated reinforcement learning, and federated meta-learning has been explored. Finally, the recent research is taking place to determine new future directions and opportunities for FL security defense mechanisms has been considered at the end of this review paper.

Abstract Emergency medicine (EM) is one of the attractive research fields in which researchers investigate their efforts to diagnose and treat unforeseen illnesses or injuries. There are many tests and observations are involved in EM. Detection of the level of consciousness is one of these observations, which can be detected using several methods. Among these methods, the automatic estimation of the Glasgow coma scale (GCS) is studied in this paper. The GCS is a medical score used to describe a patient’s level of consciousness. This type of scoring system requires medical examination that may not be available with the shortage of the medical expert. Therefore, the automatic medical calculation for a patient’s level of consciousness is highly needed. Artificial intelligence has been deployed in several applications and appears to have a high performance regarding providing automatic solutions. The main objective of this work is to introduce the edge/cloud system to improve the efficiency of the consciousness measurement through efficient local data processing. Moreover, an efficient machine learning (ML) model to predict the level of consciousness of a certain patient based on the patient’s demographic, vital signs, and laboratory tests is proposed, as well as maintaining the explainability issue using Shapley additive explanations (SHAP) that provides natural language explanation in a form that helps the medical expert to understand the final prediction. The developed ML model is validated using vital signs and laboratory tests extracted from the MIMIC III dataset, and it achieves superior performance (mean absolute error (MAE) = 0.269, mean square error (MSE) = 0.625, R2 score = 0.964). The resulting model is accurate, medically intuitive, and trustworthy.

Abstract Recently, a crucial need has grown for improving data communication for the application of roads and ad hoc networks. That is, to provide reliable and operational efficiency in data delivery and throughput. Due to the fast fragmentation and dynamic network behavior, there is an increasing demand to reach reliability in data transmission. Furthermore, the various features and the manifold of dynamic topologies in the vehicular ad hoc network raise the need to redesign the routing strategy. Thus, ensuring efficient and reliable data delivery. This paper aims to introduce a Software architecture for Road Network. The architecture is based on fog computing and aims to improve the overall performance in vehicular networks. The proposed architecture is a new routing design for the urban system to accomplish low energy consumption and operational efficiency in data delivery. The integration between the software-defined networks and fog computing platform in the proposed architecture aids to address the high rate of data transmission. Historically, this high rate negatively affected network capacity and power consumption. To prove the effectiveness of the proposed architecture, it is compared with five state-of-art algorithms published in high impact journals. The proposed architecture performance is tested based on four metrics namely packet delivery ratio, network throughput, power consumption, and routing overhead. The experimental results indicate that a 50–60% improvement in both power consumption and packet delivery ratio, while a 60–65% enhancement in network throughput and routing overhead, respectively.

Summary The problem with property valuation is that it is extremely complex. It is difficult to objectively model the pricing process or fairly estimate a property value. Many factors can contribute to this complexity such as spatial and time factors. Evaluators and researchers have been trying to model the process for centuries. Up until recently, when computer‐aided valuation systems provided better solutions in the data evaluation and real estate valuation. Nevertheless, they may suffer from low transparency, inaccuracy, and inefficiency. This work explores the ability of machine learning techniques (MLTs) in enhancing economic activities by increasing the accuracy of house price prediction. In this article, XGBoost algorithm has been integrated with outlier sum‐statistic (OS) approach. In the real estate industry, the price of property plays a crucial role in economic growth. The research attempts to predict the price of a house using MLTs. Here, the price of the property is predicted using Extreme Gradient (XG) boosting algorithm and hedonic regression pricing. Both XGBoost and hedonic pricing models use 13 variables as inputs to predict house prices. The contribution of this research lies in the practicality of using XGboost technique to predict house prices. Finally, the accuracy of the prediction algorithms is reported with XGBoosting showing the highest accuracy of 84.1% while the accuracy of the hedonic regression algorithm is 42%.