Federated Semi-Supervised Learning (FSSL) integrates Semi-Supervised Learning (SSL) with the federated framework, enabling clients to collaboratively train a global model using their local labeled and unlabeled data while preserving data privacy. Existing methods (e.g., FedMatch, FedSSL) rely on a central server to aggregate model parameters and utilize a centralized proxy dataset to guide the training process. However, these approaches face privacy risks, cause negative transfer due to data heterogeneity, and violate the lightweight design principle of federated learning. This paper proposes a decentralized framework, FedDSSL, which employs a dynamic topology structure to adaptively adjust the connection weights among clients, thereby enhancing collaborative efficiency. FedDSSL adopts a topology graph, where connection weights between clients are adaptively adjusted based on data similarity, enabling more efficient collaborative training. To replace the centralized proxy dataset, FedDSSL utilizes local self-supervised pre-training and crossclient knowledge distillation for regularization alignment. Additionally, FedDSSL introduces a distributed optimization strategy, employing multi-client collaborative validation and dynamic consistency regularization to improve the quality of pseudo-labels and model robustness. Experimental results demonstrate that FedDSSL outperforms mainstream methods in both IID and non-IID scenarios. It provides an efficient and lightweight solution for privacy-sensitive fields such as healthcare, significantly enhancing model robustness and generalization ability.

Federated learning (FL), an emerging decentralized machine learning paradigm, supports the implementation of common modeling without compromising data privacy. In practical applications, FL participants heterogeneity poses a significant challenge for FL. Firstly, clients sometimes need to design custom models for various scenarios and tasks. Secondly, client drift leads to slow convergence of the global model.Recently, knowledge distillation has emerged to address this problem by using knowledge from heterogeneous clients to improve the model’s performance. However, this approach requires the construction of a proxy dataset. And FL is usually performed with the assistance of a center, which can easily lead to trust issues and communication bottlenecks. To address these issues, this paper proposes a knowledge distillation-based FL scheme called FedDCM. Specifically, in this work, each participant maintains two models, a private model and a public model. The two models are mutual distillations, so there is no need to build proxy datasets to train teacher models. The approach allows for model heterogeneity, and each participant can have a private model of any architecture. The direct and efficient exchange of information between participants through the public model is more conducive to improving the participants’ private models than a centralized server. Experimental results demonstrate the effectiveness of FedDCM, which offers better performance compared to s the most advanced methods.

Addressing untrustworthy behavior in cross departmental business process collaboration is the focus of current research. The untrustworthiness of the business process not only leads to the misuse and leakage of business data, but also leads to cheating by the participants in the business process. This paper proposes a blockchain-based trustworthy cross-departmental business process collaboration service mechanism(TBPCS) to solve the above problems. This paper firstly maps the participants, data and data ownership of the business process to the blockchain, which prevents the business process from being tampered with. This method allows the participants in the business process to act under the constraints of a trusted environment, avoiding the illegal use of business data. Then this paper maps the business process to the blockchain in the form of smart contracts, which ensures crossdepartment business process collaboration is trustworthy and reduces the verification cost of business data. This paper innovatively proposes a rollback mechanism that supports business interruptions. The mechanism ensures that the business process is trustworthy even in abnormal situations.

The digital representation of credentials has become a necessary way in all aspects of human life, such as healthcare, education, etc. However, the current digital credentials sharing solutions tend to overlook the problem of over-disclosure. The data presentation of credentials is an all-or-nothing process, which results in the leakage of unnecessary data and threatens the privacy of the holder. In this paper, to achieve authenticated selective disclosure of credentials, we first design a hybrid storage model incorporating erasure coding (EC), where the raw data are outsourced to an off-chain distributed storage service provider while only small digest information are stored on-chain to maintain data integrity. Moreover, under the storage model, we propose an authenticated data structure (ADS) which integrates EC and the Merkle B-tree to minimize data sharing. Based on this ADS, a verifiable object (VO) can be generated, which is used to provide proof of the disclosed data without exposing the other data of the credentials. At last, we prove the security of the proposed ADS scheme and the experimental results show that, compared to a baseline solution, the proposed ADS reduces the average building and verification time, without sacrificing much of the transmission cost.

Effectively obtaining the information about the types of oil pollutants in a spill event can help determine the source of spill and formulate the plan of emergency responses. Researchers have reported the feasibility of fine-grained oil types identification using reflectance spectra, but its requirements on spectra resolution were seldom considered. To answer this question, this study examined the oil types identification accuracies using the reflection spectra with different number of bands. The reflectance spectra of various types of oil samples were collected with a high-resolution hyperspectral remote sensor, and then resampled to coarser spectral resolution. Some of the coarse-resolution spectra were built based on the central wavelengths of AVIRIS and Landsat 8 in visible bands, so as to evaluates the potential of fine-grained oil types identification using these sensors. Three kinds of machine learning algorithms were introduced as the classifier. The identification results using the reflection spectra at different resolution showed that the machine learning models could identify oil types based on high-resolution reflectance spectra. In term of the spectra at coarser resolution, the models were still able to provide accurate predictions until the number of bands was reduced to about 16 in the visible range. Therefore, the high-resolution hyperspectral sensors (e.g., AVIRIS) have the potential of fine-grained oil types identification, but that may not be accomplished by satellite-based hyperspectral sensors with less than 4 visible bands.

The rapid development of the Internet has brought new directions for improvement in e-commerce and promoted the pace of e-commerce to smart business. There are a lot of articles focusing on improving business processes efficiency, especially on business process nodes' optimization, using machine learning and even deep learning methods. Aiming at the problem of chaotic nodes' arrangement in traditional business process, this paper proposes a node optimization method, which can reconfigure business process nodes-integrating similar business nodes and decomposing complex business nodes. The algorithm proposed in this paper firstly vectorizes the process nodes to n-dimensional vectors, then uses the improved density clustering mining algorithm which is based on information entropy of clustering results to reassign the nodes. The information entropy can fairly evaluate the degree of system's confusion, so we can introduce information entropy to density clustering, then the clustering result will be evaluated and the algorithm will iteratively cluster until an optimal result is achieved. Finally, this paper conducted two experimental analysis on four business processes, successfully completed the optimal configuration of the nodes.