Federated Graph Neural Networks: Overview, Techniques, and Challenges

IEEE Transactions on Neural Networks and Learning Systems, Journal Year: 2024, Volume and Issue: 36(3), P. 4279 - 4295

Published: Feb. 8, 2024

Graph neural networks (GNNs) have attracted extensive research attention in recent years due to their capability progress with graph data and been widely used practical applications. As societies become increasingly concerned the need for privacy protection, GNNs face adapt this new normal. Besides, as clients federated learning (FL) may relationships, more powerful tools are required utilize such implicit information boost performance. This has led rapid development of emerging field (FedGNNs). promising interdisciplinary is highly challenging interested researchers grasp. The lack an insightful survey on topic further exacerbates entry difficulty. In article, we bridge gap by offering a comprehensive field. We propose 2-D taxonomy FedGNN literature: 1) main provides clear perspective integration FL analyzing how enhance training well assists GNN 2) auxiliary view FedGNNs deal heterogeneity across clients. Through discussions key ideas, challenges, limitations existing works, envision future directions that can help build robust, explainable, efficient, fair, inductive, FedGNNs.

Language: Английский

---

Traffic Pattern Classification in Smart Cities Using Deep Recurrent Neural Network

Sustainability, Journal Year: 2023, Volume and Issue: 15(19), P. 14522 - 14522

Published: Oct. 6, 2023

This paper examines the use of deep recurrent neural networks to classify traffic patterns in smart cities. We propose a novel approach pattern classification based on networks, which can effectively capture patterns' dynamic and sequential features. The proposed model combines convolutional layers extract features from data SoftMax layer patterns. Experimental results show that outperforms existing methods regarding accuracy, precision, recall, F1 score. Furthermore, we provide an depth analysis discuss implications for accurately cities with precision as high 95%. is evaluated real world dataset compared methods.

Language: Английский

---

A Survey on Heterogeneity Taxonomy, Security and Privacy Preservation in the Integration of IoT, Wireless Sensor Networks and Federated Learning

Sensors, Journal Year: 2024, Volume and Issue: 24(3), P. 968 - 968

Published: Feb. 1, 2024

Federated learning (FL) is a machine (ML) technique that enables collaborative model training without sharing raw data, making it ideal for Internet of Things (IoT) applications where data are distributed across devices and privacy concern. Wireless Sensor Networks (WSNs) play crucial role in IoT systems by collecting from the physical environment. This paper presents comprehensive survey integration FL, IoT, WSNs. It covers FL basics, strategies, types discusses WSNs various domains. The addresses challenges related to heterogeneity summarizes state-of-the-art research this area. also explores security considerations performance evaluation methodologies. outlines latest achievements potential directions emphasizes significance surveyed topics within context current technological advancements.

Language: Английский

---

Transforming Cybersecurity into Critical Energy Infrastructure: A Study on the Effectiveness of Artificial Intelligence

Systems, Journal Year: 2024, Volume and Issue: 12(5), P. 165 - 165

Published: May 5, 2024

This work explores the integration and effectiveness of artificial intelligence in improving security critical energy infrastructure, highlighting its potential to transform cybersecurity practices sector. The ability solutions detect respond cyber threats infrastructure environments was evaluated through a methodology that combines empirical analysis modeling. results indicate significant increase threat detection rate, reaching 98%, reduction incident response time by more than 70%, demonstrating identifying mitigating risks quickly accurately. In addition, implementing machine learning algorithms has allowed for early prediction failures cyber-attacks, significantly proactivity management infrastructure. study highlights importance integrating into strategies, proposing paradigmatic change increases operational efficiency strengthens resilience sustainability sector against threats.

Language: Английский

---

Federated Learning Architectures: A Performance Evaluation With Crop Yield Prediction Application

Software Practice and Experience, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 17, 2025

ABSTRACT Introduction Federated learning has become an emerging technology in data analysis for IoT applications. Methods This paper implements centralized and decentralized federated frameworks crop yield prediction based on Long Short‐Term Memory Network Gated Recurrent Unit. For learning, multiple clients one server are considered, where the exchange their model updates with that works as aggregator to build global model. framework, a collaborative network is formed among devices either using ring topology or mesh topology. In this network, each device receives from neighboring performs aggregation upgraded Results The performance of evaluated terms accuracy, precision, recall, F1‐Score, training time. experimental results show 93% accuracy achieved learning‐based frameworks. also response time can be reduced by 75% than cloud‐only framework. Conclusion Centralized architectures good loss. time, including communication both case studies, not very high, observed results. Further, no raw shared, privacy protected. Finally, future research directions use proposed.

Language: Английский

---

A contemporary survey of recent advances in federated learning: Taxonomies, applications, and challenges

Internet of Things, Journal Year: 2024, Volume and Issue: 27, P. 101251 - 101251

Published: June 15, 2024

Language: Английский

---

The Application of Machine Learning and Deep Learning in Intelligent Transportation: A Scientometric Analysis and Qualitative Review of Research Trends

Sustainability, Journal Year: 2024, Volume and Issue: 16(14), P. 5879 - 5879

Published: July 10, 2024

Machine learning (ML) and deep (DL) have become very popular in the research community for addressing complex issues intelligent transportation. This has resulted many scientific papers being published across various transportation topics over past decade. paper conducts a systematic review of literature using scientometric analysis, aiming to summarize what is already known, identify current trends, evaluate academic impacts, suggest future directions. The study provides detailed by analyzing 113 journal articles from Web Science (WoS) database. It examines growth publications time, explores collaboration patterns key contributors, such as researchers, countries, organizations, employs techniques co-authorship analysis keyword co-occurrence delve into publication clusters emerging topics. Nine sub-topics are identified qualitatively discussed. outcomes include recognizing pioneering researchers potential opportunities, identifying reliable sources information publishing new work, aiding selecting best solutions specific problems. These findings help better understand application ML DL guide policymakers editorial boards promising further development.

Language: Английский

---

Enhancing road traffic flow in sustainable cities through transformer models: Advancements and challenges

Sustainable Cities and Society, Journal Year: 2024, Volume and Issue: 116, P. 105882 - 105882

Published: Oct. 10, 2024

Language: Английский

---

ASCDNet: development of adaptive serial cascaded deep network and improved heuristic algorithm for smart transportation planning and traffic flow prediction

Journal of Ambient Intelligence and Humanized Computing, Journal Year: 2025, Volume and Issue: unknown

Published: March 18, 2025

Language: Английский

---

Utilizing Federated Learning for Enhanced Real-Time Traffic Prediction in Smart Urban Environments

International Journal of Advanced Computer Science and Applications, Journal Year: 2024, Volume and Issue: 15(2)

Published: Jan. 1, 2024

Federated Learning (FL), a crucial advancement in smart city technology, combines real-time traffic predictions with the potential to enhance urban mobility. This paper suggests novel approach prediction cities: hybrid Convolutional Neural Network-Recurrent Network (CNN-RNN) architecture. The investigation started systematic collection and preprocessing of low-resolution dataset (1.6 GB) derived from Closed Circuit Television (CCTV) camera images at significant intersections Guntur Vijayawada. has been cleaned up utilizing min-max normalization facilitate use. primary contribution this study is architecture that it develops by fusing RNN detect temporal dynamics CNN for geographic extraction characteristics. While RNN's recurrent interactions preserve hidden states sequential processing, efficiently retrieves high-level spatial information static images. Weight adjustments backpropagation are used training proposed model order aid management. Notably, implementation done Python software. reaches testing accuracy 99.8% 100th epoch, demonstrating excellent performance results discussion section. Mean Absolute Error (MAE) results, which show 4.5% improvement over existing methods like Long Short Term Memory (LSTM), Support Vector Machine (SVM), Sparse Auto Encoder (SAE), Gated Recurrent Unit (GRU), illustrate efficacy model. demonstrates how well complex patterns may be represented model, yielding precise crowded metropolitan settings. A new era more effective forecasts about begin, thanks CNN-RNN architecture, validated combined strengths FL, CNN, as overall outcomes.

Language: Английский

---