Split Learning (SL) is a distributed privacy-preserving learning methodology designed to address the challenges associated with the deployment of large-scale deep neural networks on medical devices, while simultaneously safeguarding the privacy of medical data. However, both the forward and backward propagation of the model require communication between the medical devices and high-performance servers, resulting in significant communication overhead and high latency. In this paper, to reduce the communication overhead from the client to the server during forward propagation, we propose an autoencoder layer based on attention mechanisms and triple compression. To reduce the communication overhead from the server to the client during backward propagation, we proposed an average loss threshold algorithm to decrease the frequency of client updates. Compared to the original Split Learning algorithm, after incorporating the method proposed in this paper, the communication overhead during forward propagation decreased by an average of 93%, and during backward propagation, it decreased by an average of 96%. The total communication overhead decreased by an average of 95%. The model’s accuracy loss was between 0% and 1%. In terms of communication compression, compared to the SOTA SL–BSL, the overall communication overhead was reduced by an average of 92%.

The integration of Artificial Intelligence (AI) in the healthcare sector has marked significant advancements, and Federated Learning (FL) has further facilitated the amalgamation of Federated Medical Imaging. However, this integration has also sparked concerns regarding data privacy. Incorporating Differential Privacy (DP) into gradients effectively mitigates privacy leaks but at the cost of impacting model accuracy. Current research delves into DP within FL, with a focus on strategies for privacy budget allocation and noise addition. Nevertheless, the dynamic privacy requirements and resource optimization for actual medical applications are often overlooked, leading to resource wastage. This study introduces an innovative algorithm based on gradient component for adaptive noise scale optimization and privacy budget allocation, thereby enhancing privacy management while maintaining model accuracy. Our findings reveal that, compared to traditional DP techniques, our approach achieves an average accuracy improvement of 3.47% in RSNA-ICH Acc and an enhancement of up to 172.33% in Dice score for Prostate MRI Dice under various privacy budgets, demonstrating the substantial efficacy of our method in the domain of federated medical imaging.

Federated Learning (FL), a distributed learning paradigm optimizing communication costs and enhancing privacy by uploading gradients instead of raw data, now confronts security challenges. It is particularly vulnerable to Byzantine poisoning attacks and potential privacy breaches via inference attacks. While homomorphic encryption and secure multi-party computation have been employed to design robust FL mechanisms, these predominantly rely on Euclidean distance or median-based metrics and often fall short in comprehensively defending against advanced poisoning attacks, such as adaptive attacks. Addressing this issue, our study introduces “Split-Aggregation", a lightweight privacy-preserving FL solution capable of withstanding adaptive attacks. This method maintains a computational complexity ofO(dkN+k3) and a communication overhead ofO(dN), performing comparably to FedAvg whenk= 10. Here,drepresents the gradient dimension,Nthe number of users, andkthe rank chosen during random singular value decomposition. Additionally, we utilize adaptive weight coefficients to mitigate gradient descent issues in honest users caused by non-independent and identically distributed (Non-IID) data. The proposed method’s security and robustness are theoretically proven, with its complexity thoroughly analyzed. Experimental results demonstrate that atk= 10, this method surpasses the top-1 accuracy of current state-of-the-art robust privacy-preserving FL approaches. Moreover, opting for a smallerksignificantly boosts efficiency with only marginal compromises in accuracy.

ABSTRACT With the development of modern Internet and mobile networks, there is an increasing need for collaborative privacy‐preserving applications. Secure multi‐party computation (SMPC) gives a general solution to these applications and has become a hot topic. Yao's garbled circuit approach is a leading method in designing protocols for secure two‐party computation (2PC), which is a very important base in SMPC. However, there are only few protocols obtaining the fairness of secure 2PC, and only one of them was constructed within the standard simulation framework but with very low efficiency. In this paper, we propose an efficient fair secure Yao's garbled circuit protocol within the universally composable (UC) framework. By comparing with all other fair secure Yao's protocols, our new protocol enjoys three advantages. First, our protocol is more efficient than any other fair secure Yao's protocols within the standard simulation framework. Second, our protocol is the first fair UC‐secure Yao's garbled circuit protocol, so it is more secure than other fair Yao's protocols. Third, there does not require any third party involved in our protocol; thus, it is very suitable for many applications. Copyright © 2013 John Wiley & Sons, Ltd.

P2P Networks are self-organized and distributed. Efficient nodes risk assessment is one of the key factors for high quality resource exchanging. Most assessment methods based on trust or reputation have some remarkable drawbacks. For example, some methods impose too many restrictions to the samples, and many methods can’t identify the malicious recommendations, which result in that the final results are not convincible and credible. To solve these problems, we propose a novel risk assessment method based on grey theory. In our scheme, the communication nodes’ incomplete information state is described as several key attributes. Original data of these attributes is collected using taste concourse method to avoid malicious recommendation. The analysis and computing example shows this scheme is an efficient incomplete information nodes risk assessment method in P2P networks

The secure interaction between two or more administrative domains is a major concern. IRBAC2000 is a model that quickly establishes a flexible policy for dynamic role translation from foreign domains to local. A-IRBAC2000 mode utilizes RBAC to manage dynamic role translation between foreign and local domains. We will see that these mechanisms have significant shortcomings. We propose an improved administrative usage control model named AUCON to overcome the weakness of previous models. AUCON provides administrates user-role assignment for local and foreign domain with unified method. It provides flexible enough mechanism to distinguish users of foreign and local domain and can enforce more strict control for foreign user. While retaining the advantage of traditional RBAC model, AUCON model is being implemented in experiment system