Federated Learning (FL) has emerged as a promising methodology for collaboratively training machine learning models on decentralized devices. Notwithstanding, the effective synchronization and consolidation of model updates originating from diverse devices, in conjunction with the appropriate configuration of network topologies, persist as crucial obstacles. This paper provides a comprehensive analysis of the current techniques and methodologies utilized in the synchronization, consolidation, and network topologies of Federated Learning. The present study explores diverse synchronization strategies utilized for the purpose of coordinating model updates from geographically distributed cross-silo edge nodes. The study takes into account several factors, including communication efficiency and privacy preservation. This study delves into the intricacies of model consolidation techniques, such as weighted and personalized aggregation methods, to evaluate their efficacy in consolidation of local model updates into a global model, while taking into consideration statistical heterogeneity and resource constraints. In addition, an examination is conducted on the importance of network topologies in Federated Learning (FL), taking into account their influence on communication efficacy, confidentiality, expandability, resilience, and resource allocation. The survey assesses and contrasts the efficacies and constraints of extant methodologies, discerns deficiencies in present investigations, and provides insights for future progressions. The objective of this survey is to provide a thorough examination of FL synchronization, consolidation, and network topologies, with the intention of offering a valuable reference for individuals engaged in Federated Learning, including researchers, practitioners, and stakeholders. This survey aims to support the advancement of more effective and resilient FL systems.

Recent terrorist attacks have emerged as a formidable menace to global peace and security, giving rise to an acute humanitarian and economic crisis characterized by the loss of numerous lives and the incurring of substantial financial damages. In response to this pressing concern, the scholarly community has introduced a range of AI-driven predictive analytics methodologies as prospective instruments in the fight against terrorism. Machine Learning (ML) techniques, commonly employed in counter-terrorism efforts, face the formidable challenge of accurately forecasting terrorism activities due to the escalating complexity and voluminous nature of the underlying data. In this context, we present a solution in the form of an optimal weighted voting ensemble classifier, specially tailored for the classification of weapon types in terrorist attacks. Leveraging the computational capabilities of the Particle Swarm Optimization (PSO) algorithm, we determine the optimal weight assignments for the base learners, namely Random Forest and Xtreme Gradient Boosted Machines. Additionally, we developed several machine learning model crafted to predict the likelihood of casualties in the aftermath of a terrorist incident. Rigorous validation of these models is carried out utilizing the publicly available Global Terrorism Database (GTD), necessitating meticulous data preprocessing to rectify anomalies and address class imbalances, a task effectively accomplished through the application of the Synthetic Minority Over-sampling Technique (SMOTE). To underscore the effectiveness of our proposed model, we conduct an extensive comparative analysis, bench-marking it against state-of-the-art machine learning models. The comprehensive experimental results unequivocally demonstrate the superiority of our models, as they consistently achieve the lowest error rates, thereby highlighting their enhanced generalizability and performance. The findings derived from our study furnish invaluable insights to empower counter-terrorism agencies, facilitating data-driven decision-making and the proactive implementation of measures to mitigate and counteract the scourge of terrorist acts.

Intrusion Detection Systems (IDSs) play a crucial role in safeguarding network infrastructures from cyber threats and ensuring the integrity of highly sensitive data. Conventional IDS technologies, although successful in achieving high levels of accuracy, frequently encounter substantial model bias. This bias is primarily caused by imbalances in the data and the lack of relevance of certain features. This study aims to tackle these challenges by proposing an advanced machine learning (ML) based IDS that minimizes misclassification errors and corrects model bias. As a result, the predictive accuracy and generalizability of the IDS are significantly improved. The proposed system employs advanced feature selection techniques, such as Recursive Feature Elimination (RFE), sequential feature selection (SFS), and statistical feature selection, to refine the input feature set and minimize the impact of non-predictive attributes. In addition, this work incorporates data resampling methods such as Synthetic Minority Oversampling Technique and Edited Nearest Neighbor (SMOTE_ENN), Adaptive Synthetic Sampling (ADASYN), and Synthetic Minority Oversampling Technique–Tomek Links (SMOTE_Tomek) to address class imbalance and improve the accuracy of the model. The experimental results indicate that our proposed model, especially when utilizing the random forest (RF) algorithm, surpasses existing models regarding accuracy, precision, recall, and F Score across different data resampling methods. Using the ADASYN resampling method, the RF model achieves an accuracy of 99.9985% for botnet attacks and 99.9777% for Man-in-the-Middle (MITM) attacks, demonstrating the effectiveness of our approach in dealing with imbalanced data distributions. This research not only improves the abilities of IDS to identify botnet and MITM attacks but also provides a scalable and efficient solution that can be used in other areas where data imbalance is a recurring problem. This work has implications beyond IDS, offering valuable insights into using ML techniques in complex real-world scenarios.

Document Clustering has attracted the interest of many researchers who have created several solutions to this problem by combining different techniques, models, and algorithms. While famous for its simplicity, the most commonly used algorithm, K-Means, suffers from issues such as finding the optimal value for k and random initialization of the centroids. In this paper, we propose a hybrid methodology combining K-Means++ with the metaheuristic algorithm PSO to overcome the challenges of both these algorithms. K-Means++ is a smart initialization technique that selects clusters based on probability function so that they are further away from each other. Two methods, Elbow Analysis, and Silhouette Score, have been used to find the optimal value of k. The research also attempts to assess the results of the best-performing feature extraction techniques and create a combined approach using their average. Although the combined approach did not outperform the existing methods, it paved the way for further advancement and exploration of merging feature extraction techniques. The experiments were performed on four datasets: 20 Newsgroups, Reuters, WebKB, and Wine, and the results were evaluated using three evaluation metrics: Purity, Silhouette Score, and Jaccard Index. The proposed approach achieved a Purity of 0.95, 0.91. 0.86, 0.89, the Silhouette Score was measured to be 0.5, 0.05, 0.05, 0.59, and the Jaccard Index was reported to be 0.04, 0.04, 0.05, 0.14 on Reuters, WebKB, 20 Newsgroups, and Wine datasets, respectively.