The emerging integration of Internet of Things (IoT) and AI has unlocked numerous opportunities for innovation across diverse industries. However, growing privacy concerns and data isolation issues have inhibited this promising advancement. Unfortunately, traditional centralized Machine Learning (ML) methods have demonstrated their limitations in addressing these hurdles. In response to this ever-evolving landscape, Federated Learning (FL) has surfaced as a cutting-edge ML paradigm, enabling collaborative training across decentralized devices. FL allows users to jointly construct AI models without sharing their local raw data, ensuring data privacy, network scalability, and minimal data transfer. One essential aspect of FL revolves around proficient knowledge aggregation within a heterogeneous environment. Yet, the inherent characteristics of FL have amplified the complexity of its practical implementation compared to centralized ML. This survey delves into three prominent clusters of FL research contributions: personalization, optimization, and robustness. The objective is to provide a well-structured and fine-grained classification scheme related to these research areas through a unique methodology for selecting related work. Unlike other survey papers, we employed a hybrid approach that amalgamates bibliometric analysis and systematic scrutinizing to find the most influential work in the literature. Therefore, we examine challenges and contemporary techniques related to heterogeneity, efficiency, security, and privacy. Another valuable asset of this study is its comprehensive coverage of FL aggregation strategies, encompassing architectural features, synchronization methods, and several federation motivations. To further enrich our investigation, we provide practical insights into evaluating novel FL proposals and conduct experiments to assess and compare aggregation methods under IID and non-IID data distributions. Finally, we present a compelling set of research avenues that call for further exploration to open up a treasure of advancement.
24 November 1998 in IEE Proceedings - Control Theory and Applications
Fuzzy reasoning models are model-free estimators of control systems, which makes them very powerful tools in control applications. Their performances depend on several factors such as the completeness of the rule base, the subjective definitions of the fuzzy subset parameters or fuzzy partitioning, the fuzzy implication operator and the defuzzification method. In most cases these factors are decided by the subjective experience of the human expert. The authors attempt to improve the fuzzy systems performances by means of a genetic-based learning mechanism. They propose a new methodology based upon two genetic algorithm (GA) loops; the inner GA loop enables the generation of an optimal set of parameters of the fuzzy reasoning matrix based upon their selection by particular optimal fuzzy partitioning generated by an outer GA loop. Simulation of typical control problems shows the effectiveness and high performance of such a method. The effects of GA parameter settings on convergence and total performance are also discussed.
During the last decade Wireless Sensor Networks and IoT have taken an overwhelming place in engineering and environmental applications. Thus, with growing interest in large size and complex networks, some critical challenges have to be taken care of such as energy consumption and node localization, especially when dealing with real-time applications. Such factors affect strongly the system performance in sensitive fields like e-health or military applications. The primary purpose is to determine the location of the sensor node that triggers the event. Energy consumption of each node is a serious problem for WSN when it comes to extend the life time of the whole network. In this paper, we propose a multicriteria optimization scheme based on a bio-inspired algorithm called Harris Hawks Optimization Algorithm (HHOA). It is shown that the proposed paradigm is capable of increasing the localization rate, as well as minimizing the energy consumption of the nodes. HHOA populations are able to share information in a multi-agent fashion to compute the trigger's position. Developing this algorithm on a huge WSN with millions of nodes reveals a particularly high performance. To assess this, several experiments in different scenarios are carried out in a decentralized environment of WSN. Finally, a comparative study is carried out also to some recent bio-inspired algorithms.
In the retail industry, Supply Chain Management (SCM) holds significant importance as it ensures the efficient movement of goods from suppliers to customers. In this intricate and fast-paced environment, the availability of accurate information and data is crucial. The purpose of this paper is to develop a framework that enhances forecasting accuracy and efficiency in Supply Chain (SC) operations within the retail industry. By analyzing the latest research and advancements in the field, thi s paper seeks to contribute valuable insights into the potential of deep learning for SCM. The ultimate goal is to provide retailers with a reliable tool that empowers them to make informed decisions based on accurate predictions, thereby optimizing their SC operations and better meeting customer demands in the dynamic retail landscape. Furthermore, DLSTM-SCM framework updates dynamically the deployed Long Short-Term Memory (LSTM) models to predict sales for the upcoming day sales, utilizing historical sales data and incorporating statistical features such as lagging and shifting to enhance forecasting precision. The efficacy of DLSTM-SCM is demonstrated through its performance on real benchmarks, where it yielded significant improvements compared to existing methods.
The rise of social networks has increased user interactions but also led to a surge in fake accounts, threatening privacy and security. These fraudulent profiles undermine trust and facilitate misinformation, making their detection crucial. This study proposes an improved machine learning approach using the Negative Selection Algorithm (NSA) to enhance fake account identification. Inspired by the human immune system, NSA helps distinguish between genuine and fraudulent accounts. Additionally, K-Means clustering refines the training dataset by removing anomalies, improving classification accuracy. Experimental results show promising improvements, contributing to more effective fake account detection and a safer online environment.
Due to the increasing frequency of forest fires and their rapid spread, early detection is critical for effective containment and mitigation. This paper proposes a cost-effective edge-based forest fire detection system that receives images from multiple sources (terrestrial cameras and UAVs) to predict and alert authorities of potential forest fires. The model of the system is built using transfer learning with MobileNetv2 on a realistic and diverse dataset, resulting in a lightweight CNN model that is further optimized by using quantization to reduce its size and improve the inference speed. The proposed model is deployed on an 8-node Raspberry Pi cluster, using Slurm and MPI to manage cluster task scheduling and parallel processing. The proposed system achieves 99.21% accuracy, precision, recall, and F1 score on a realistic test dataset containing vague real-world scenarios such as fog and sunset conditions, with an inference speed of 69 frames per second. These results, along with the system's autonomous and offline operation, cost effectiveness, power efficiency, and scalability, make it ideal for real-time forest fire monitoring at the edge, even in off-grid and remote areas.