The rapid expansion of Internet of Things (IoT) devices and advancements in mobile computing have underscored the significance of edge computing in addressing data processing and analytics challenges at the network. Federated Learning (FL) emerges as a strategic response, enabling collaborative model training across multiple participants without necessitating the sharing of raw data, thus preserving privacy and effectively leveraging decentralized data. However, FL's architecture presents distinct security challenges, particularly in safeguarding data and models against unauthorized access and manipulation. This paper presents a FL framework that integrates Attribute-Based Encryption (ABE) to provide fine-grained data access control while ensuring data security. In addition, we employ device fingerprints to enhance user authentication and device security, ensuring that only verified devices can participate in the model training process, thereby minimizing the risks of malicious access and data breaches. This paper delineates our contributions towards enhancing FL security, including the deployment of encryption methodologies, access management, and a comprehensive discussion on security mechanisms.

In the rapidly evolving field of machine learning, the adoption of traditional approaches often encounters limitations, such as increased computational costs and the challenge of catastrophic forgetting, particularly when models undergo retraining with new datasets. This issue is especially pronounced in environments that require the ability to swiftly adapt to changing data landscapes. Continual learning emerges as a pivotal solution to these challenges, empowering models to assimilate new information while preserving the knowledge acquired from previous learning phases. Despite its benefits, the continual learning process's inherent need to retain prior knowledge introduces a potential risk for information leakage. Addressing these challenges, we propose a Federated Continual Learning (FCL) framework with a multi-head neural network model. This approach blends the privacy-preserving capabilities of Federated Learning (FL) with the adaptability of continual learning, ensuring both data privacy and continuous learning in edge computing environments. Moreover, this framework enhances our approach to adversarial training, as the constant influx of diverse and complex training data allows the model to improve its understanding and adaptability, thereby strengthening its defenses against adversarial threats. Our system features a architecture with dedicated fully-connected layers for each task, ensuring that unique features pertinent to each task are accurately captured and preserved over the model's lifetime. Data undergoes processing through these task-specific layers before a final label is determined, based on the highest prediction value. This method exploits the model's full range of knowledge, significantly boosting prediction accuracy. We have conducted thorough evaluations of our FCL framework on two benchmark datasets, MNIST and CIFAR-10, with the results clearly validating the effectiveness of our approach.

While recent years have witnessed the advancement in big data and artificial intelligence, it is of much importance to safeguard data privacy and security. As an innovative approach, federated learning (FL) addresses these concerns by facilitating collaborative model training across distributed data sources without transferring raw data. However, the challenges of robust security and privacy across decentralized networks catch significant attention in dealing with the distributed data in FL. In this paper, we conduct an extensive survey of the security and privacy issues prevalent in FL, underscoring the vulnerability of communication links and the potential for cyber threats. We delve into various defensive strategies to mitigate these risks, explore the applications of FL across different sectors, and propose research directions. We identify the intricate security challenges that arise within the FL frameworks, aiming to contribute to the development of secure and efficient FL systems.

Recently, as big data and AI technology advance, data privacy and security are increasingly critical. Federated Learning (FL) has become a key solution in machine learning to address these concerns. In this paper, we present a secure and lightweight FL scheme. It employs masking and Secret Sharing (SS) to securely aggregate data from distributed clients, thereby reducing the demands of model training on system resources. The scheme also computes data similarity among clients to evaluate each client's contribution, defending against challenges posed by malicious clients. This approach safeguards privacy, facilitates accurate model updates, and addresses the challenges of limited resources in edge computing environments. We subjected our framework to rigorous validation using MNIST datasets. Experimental outcomes unequivocally substantiate the efficacy of our proposed methodology.

While centralized servers pose a risk of being a single point of failure, decentralized approaches like blockchain offer a compelling solution by implementing a consensus mechanism among multiple entities. Merging distributed computing with cryptographic techniques, decentralized technologies introduce a novel computing paradigm. Blockchain ensures secure, transparent, and tamper-proof data management by validating and recording transactions via consensus across network nodes. Federated Learning (FL), as a distributed machine learning framework, enables participants to collaboratively train models while safeguarding data privacy by avoiding direct raw data exchange. Despite the growing interest in decentralized methods, their application in FL remains underexplored. This paper presents a thorough investigation into Blockchain-based FL (BCFL), spotlighting the synergy between blockchain's security features and FL's privacy-preserving model training capabilities. First, we present the taxonomy of BCFL from three aspects, including decentralized, separate networks, and reputation-based architectures. Then, we summarize the general architecture of BCFL systems, providing a comprehensive perspective on FL architectures informed by blockchain. Afterward, we analyze the application of BCFL in healthcare, IoT, and other privacy-sensitive areas. Finally, we identify future research directions of BCFL.

Federated Learning (FL) enables model training directly on edge devices, thereby enhancing privacy by keeping data local. Despite advances in FL, the dynamic nature of data requires models to continuously adapt without forgetting previously acquired knowledge, thus leading to the adoption of Federated Continual Learning (FCL). However, this prior knowledge must be persistently stored in model parameters to mitigate catastrophic forgetting, which increases the risk of data or sensitive information exposure. Therefore, to bolster the privacy of local information, Differential Privacy (DP) is implemented as a privacy protection technique in FL, and its effectiveness is evaluated. Using the MNIST dataset, we assess the impact of noise on model accuracy under both Independent and Identically Distributed (IID) and Non-Independent and Identically Distributed (Non-IID) data distributions. We demonstrate that the proposed method not only adapts to changes in the operational environment but also significantly mitigates the risk of catastrophic forgetting.

Federated Learning (FL) and the Internet of Things (IoT) have revolutionized data processing and analysis, overcoming the traditional limitations of cloud computing. However, traditional machine learning strategies lead to increased costs and catastrophic forgetting due to model retraining with new datasets. Continual learning has been proposed to counter this, enabling models to adapt to new data while preserving previous knowledge, which is beneficial for dynamic edge environments. Despite these advantages, the retention of previous knowledge during the continual learning process may lead to the information leakage. To address the inherent challenges of multitask scenarios, we present a Federated Continual Learning (FCL) framework that integrates the privacy-preserving benefits of Federated Learning (FL) into a continual learning system, ensuring both continual learning and privacy preservation in edge computing data processing and analysis. Specifically, our architecture introduces dedicated fully-connected layers for each task. This architecture ensures that distinctive features pertinent to each task are not only captured but also preserved throughout the model’s lifespan. Within our framework, data is processed via task-specific layers. Subsequently, the final label is determined by associating it with the paramount prediction value, thus capitalizing on the model’s comprehensive knowledge reservoir to bolster prediction accuracy. We subjected our FCL framework to rigorous validation using two benchmark datasets: MNIST and CFAR-10. Experimental outcomes unequivocally substantiate the efficacy of our proposed methodology.

Federated Learning (FL) and the Internet of Things (IoT) have transformed data processing and analysis, overcoming traditional cloud computing limitations. However, challenges such as catastrophic forgetting in continuous training scenarios arise. To address these, we propose an FL framework that supports continual learning while enhancing system security. We preserve critical knowledge through the incorporation of Knowledge Distillation (KD), addressing the issue of catastrophic forgetting. In addition, we have integrated encryption techniques to secure the updated parameters of clients from potential threats posed by attackers.

Internet service has enabled digital contents to be shared faster and easier, but on the other side it raised an issue of illegal copy of the digital contents. Public key encryption schemes solve this issue partially. However, there is still a weak point that the secret key is not completely protected; that is, public key encryption schemes suffer from illegal copy of secret keys (the key-cloning problem). In this paper, first, we discuss the usability of terminal fingerprints for key-cloning protection. Next, we propose a hybrid encryption scheme using terminal fingerprints to protect the secret keys from the key-cloning. Based on an assumption that the terminal fingerprint is unchangeable and unextractable even by the user of the terminal, our hybrid encryption scheme can be effectively used as a method of the key-cloning protection. Then, we instantiate our hybrid encryption scheme as a combination of the attribute-based encryption scheme and the RSA encryption scheme; the attribute-based encryption scheme functions as a mechanism for authentication of user’s attributes, and the RSA encryption scheme functions as a mechanism for authentication of a terminal device. Terminal fingerprint that is a feature of the device is used to generate a secret key of the RSA encryption.

Internet services make sharing digital contents faster and easier but raise an issue of illegal copying and distribution of those digital contents at the same time. A lot of public key encryption schemes solve this issue. However, the secret key is not completely protected i.e. these kinds of encryption methods do not prevent illegal copying and distribution of secret keys. In this paper, we propose a hybrid encryption scheme that employ terminal fingerprints. This scheme is a template to avoid such misuse of secret keys, and can be applied to, for example, attribute-based encryption schemes. There terminal fingerprint information is used to create a second encryption key and secret key. Since the terminal fingerprint is assumed to be unchangeable and unknowable, we ensure that our secret keys are valid in the terminal where such secret keys were created.