The current COVID-19 pandemic threatens human life, health, and productivity. AI plays an essential role in COVID-19 case classification as we can apply machine learning models on COVID-19 case data to predict infectious cases and recovery rates using chest x-ray. Accessing patient’s private data violates patient privacy and traditional machine learning model requires accessing or transferring whole data to train the model. In recent years, there has been increasing interest in federated machine learning, as it provides an effective solution for data privacy, centralized computation, and high computation power. In this paper, we studied the efficacy of federated learning versus traditional learning by developing two machine learning models (a federated learning model and a traditional machine learning model)using Keras and TensorFlow federated, we used a descriptive dataset and chest x-ray (CXR) images from COVID-19 patients. During the model training stage, we tried to identify which factors affect model prediction accuracy and loss like activation function, model optimizer, learning rate, number of rounds, and data Size, we kept recording and plotting the model loss and prediction accuracy per each training round, to identify which factors affect the model performance, and we found that softmax activation function and SGD optimizer give better prediction accuracy and loss, changing the number of rounds and learning rate has slightly effect on model prediction accuracy and prediction loss but increasing the data size did not have any effect on model prediction accuracy and prediction loss. finally, we build a comparison between the proposed models’ loss, accuracy, and performance speed, the results demonstrate that the federated machine learning model has a better prediction accuracy and loss but higher performance time than the traditional machine learning model.

The industry has recognized the risk of cyber-attacks targeting to the advanced metering infrastructure (AMI). A potential adversary can modify or inject malicious data, and can perform security attacks over an insecure network. Also, the network operators at intermediate devices can reveal private information, such as the identity of the individual home and metering data units, to the third-party. Existing schemes generate large overheads and also do not ensure the secure delivery of correct and accurate metering data to all AMI entities, including data concentrator at the utility and the billing center. In this paper, we propose a secure and privacy-preserving data aggregation scheme based on additive homomorphic encryption and proxy re-encryption operations in the Paillier cryptosystem. The scheme can aggregate metering data without revealing the actual individual information (identity and energy usage) to intermediate entities or to any third-party, hence, resolves identity and related data theft attacks. Moreover, we propose a scalable algorithm to detect malicious metering data injected by the adversary. The proposed scheme protects the system against man-in-the-middle, replay, and impersonation attacks, and also maintains message integrity and undeniability. Our performance analysis shows that the scheme generates manageable computation , communication, and storage overheads and has efficient execution time suitable for AMI networks.

No abstract

Abstract-Vehicular communications networks are envisioned for the access to drive-thru Internet and IP-based infotainment applications. These services are supported by roadside access routers (ARs) that connect vehicular ad hoc networks (VANETs) to external IP networks. However, VANETs suffer from asymmetric links due to variable transmission ranges caused by mobility, obstacles, and dissimilar transmission power, which make it difficult to maintain the bidirectional connections and to provide the IP mobility required by most IP applications. Moreover, vehicular mobility results in short-lived connections to the AR, affecting the availability of IP services in VANETs. In this paper, we study the secure and timely handover of IP services in an asymmetric VANET and propose a multihop-authenticated Proxy Mobile IP (MA-PMIP) scheme. MA-PMIP provides an enhanced IP mobility scheme over infrastructure-to-vehicle-to-vehicle (I2V2V) communications that uses location and road traffic information. The MA-PMIP also reacts, depending on the bidirectionality of links, to improve availability of IP services. Moreover, our scheme ensures that the handover signaling is authenticated when V2V paths are employed to reach the infrastructure so that possible attacks are mitigated without affecting the performance of the ongoing sessions. Both analysis and extensive simulations in OMNeT++ are conducted, and the results demonstrate that the MA-PMIP improves service availability and provides secure seamless access to IP applications in asymmetric VANETs.Index Terms-Asymmetric links, infrastructure-to-vehicle-tovehicle (I2V2V), IP mobility, multihop networks, mutual authentication, Proxy Mobile IP (PMIP), vehicular ad hoc network (VANET).

No abstract