23 research outputs found
Utilizing Blockchain for Distributed Machine Learning based Intrusion Detection in Internet of Things
No full textInstance-Based Lossless Summarization of Knowledge Graph With Optimized Triples and Corrections (IBA-OTC)
No full textKnowledge graph (KG) summarization facilitates efficient information retrieval for exploring complex structural data. For fast information retrieval, it requires processing on redundant data. However, it necessitates the completion of information in a summary graph. It also saves computational time during data retrieval, storage space, in-memory visualization, and preserving structure after summarization. State-of-the-art approaches summarize a given KG by preserving its structure at the cost of information loss. Additionally, the approaches not preserving the underlying structure, compromise the summarization ratio by focusing only on the compression of specific regions. In this way, these approaches either miss preserving the original facts or the wrong prediction of inferred information. To solve these problems, we present a novel framework for generating a lossless summary by preserving the structure through super signatures and their corresponding corrections. The proposed approach summarizes only the naturally overlapped instances while maintaining its information and preserving the underlying Resource Description Framework RDF graph. The resultant summary is composed of triples with positive, negative, and star corrections that are optimized by the smart calling of two novel functions namely merge and disperse . To evaluate the effectiveness of our proposed approach, we perform experiments on nine publicly available real-world knowledge graphs and obtain a better summarization ratio than state-of-the-art approaches by a margin of 10% to 30% with achieving its completeness, correctness, and compactness. In this way, the retrieval of common events and groups by queries is accelerated in the resultant graph
Networked Federated Meta-Learning Over Extending Graphs
No full textDistributed and collaborative machine learning over emerging Internet of Things (IoT) networks is complicated by resource constraints, device, and data heterogeneity, and the need for personalized models that cater to the individual needs of each network device. This complexity becomes even more pronounced when new devices are added to a system that must rapidly adapt to personalized models. Along these lines, we propose a networked federated meta-learning (NF-ML) algorithm that utilizes meta-learning and underlying shared structures across the network to enable fast and personalized model adaptation of newly added network devices. The NF-ML algorithm learns two sets of model parameters for each device in a distributed manner, with devices communicating only with their immediate neighbors. One set of parameters is personalized for the device-specific task, whereas the other is a generic parameter set learned via peer-to-peer communication. The performance of the proposed NF-ML algorithm was validated using both synthetic and real-world data, and the results show that it adapts to new tasks in just a few epochs, using as little as 10% of the available data, significantly outperforming traditional federated learning methods.</p
Region-specific reliable channel estimation in RIS-enabled wireless communications via clustered federated learning
No full textMachine learning (ML)-based downlink channel estimation for reconfigurable intelligent surface (RIS)-assisted communication faces challenges such as handling channel variations, high communication overhead of centralized learning (CL), and vulnerability to malicious users. We propose a novel approach integrating blockchain to enhance security by verifying registered users, autoencoder (AE)-based clustering to identify regions within the cell, and clustered federated learning (CFL) to ensure good channel estimation performance while minimizing communication and energy overhead. Simulations show that the proposed clustering-based scheme achieves estimation performance comparable to CL while significantly reducing communication and energy overhead.</p
Channel Estimation in RIS-Aided Heterogeneous Wireless Networks via Federated Learning
No full textDownlink channel estimation in reconfigurable intelligent surface (RIS)-assisted communication systems employing federated learning (FL) is challenging due to communication/ computational overhead, users heterogeneity, and vulnerability to malicious users. This letter proposes a novel methodology integrating principal component analysis (PCA)-based clustering with FL, tailored for heterogeneous users. The approach effectively identifies regions and users within the cell while minimizing communication/computational overhead associated with clustering, resulting in accurate, resource-efficient, and secure channel estimation. Simulation results demonstrate that the proposed FL strategy achieves estimation performance comparable to the conventional methods while significantly reducing the communication overhead, enhancing the system security, and handling heterogeneous users.</p
Blockchain-based Secure Delivery of Medical Supplies Using Drones
No full textThe advantages provided by the drones with regards to three dimensional mobility and ease of deployment makes them a viable candidate for 5G and beyond (B5G) networks. Significant amount of research has been conducted on the aspect of networking for using drones as base stations to provide different services. In this work, we deviate from the traditional use of drones to provide connectivity and explore the delivery of products through drones in the context of maintaining social distancing. However, drone delivery process for critical applications such as delivering medical supplies is vulnerable to attacks such as impersonation attacks and eavesdropping. The security of drone operation is important to save the users from any breaches that can lead to financial and physical losses. To cope with these security issues and to make the delivery process transparent, we propose a blockchain-based drone delivery system that registers and authenticates the participating entities including products (medical supplies), warehouse (medical centers) and drones. To this end, we utilize Ethereum platform for implementation of blockchain and smart contract and we present an analysis of different factors that influence the authentication process in terms of time and the number of transactions. Furthermore, to make the communication of a drone with command and control center more secure and robust, we use machine learning (ML)-based intrusion prevention system to detect any impersonation attacks with an accuracy of 97%
Computationally-Efficient Structural Health Monitoring using Graph Signal Processing
No full textStructural health monitoring (SHM) of bridges is crucial for ensuring safety and long-term durability, however, standard damage-detection algorithms are computationally intensive. This article proposes a computationally efficient algorithm based on graph signal processing (GSP) to leverage the underlying network structure in the data. Under the assumption that damages impact both spatial and temporal structures of the sensor data, the algorithm combines spatial and temporal information from accelerometers by computing the smoothness of graph signals expanded along time. The Kullback-Leibler (KL) divergence is used as dissimilarity metric to distinguish between healthy condition and presence of a damage, while Tukey's method for outliers removal and sequential detection via exponential weighted moving average (EWMA) are then employed for performance improvement. The proposed GSP-based SHM system is appealing in terms of simplicity and low-complexity and is also suitable for real-time monitoring. The effectiveness in terms of detection performance is validated both on synthetically generated data and real-world measurements.</p
Data-Driven Classifiers for Early Meal Detection Using ECG
No full textThis study investigates the potential of the electrocardiogram (ECG) to perform early meal detection, which is critical for developing a fully-functional automatic artificial pancreas. The study was conducted in a group of healthy subjects with different ages and genders. Two classifiers were trained: one based on neural networks (NNs) and working on features extracted from the signals; one based on convolutional NNs (CNNs) and working directly on raw data. During the test phase, both classifiers correctly detected all the meals, with the CNN outperforming the NN in terms of misdetected meals (MMs) and detection time (DT). Reliable meal onset detection with short detection time has significant practical implications: it reduces the risk of postprandial hyperglycemia and hypoglycemia, it reduces the mental burden of meal documentation for patients and related stress.Data-Driven Classifiers for Early Meal Detection Using ECGacceptedVersio
