Hybrid Data Processing by Combining Machine Learning, Expert, Safety and Security

The goal of this Special Issue is to promote hybrid data processing by combining machine learning with experts' input, data safety, and security. AI technology and machine learning technology are developing rapidly. Data contain important information that can advance human knowledge and enhance AI capabilities. Meanwhile, requirements for data mining and data processing are expanding. Machine learning and deep learning may achieve excellent results, but in some cases, a balance can be reached by involving experienced experts to save resources and improve outcomes. In mining and analyzing data, the issues of data safety, data security, and data privacy also need to be suitably considered. This Special Issue presents ten rigorously reviewed manuscripts that study how to integrate hybrid data intelligence with experts' input, expert systems, safety, and security through decentralized reputation systems, blockchain technology, linkable ring signatures, collaborative filtering, contrastive learning, graph neural networks, feature selection, sample imbalance, few-shot learning, contrastive learning, knowledge graphs, transfer learning, dynamic Gaussian Bayesian networks, the Manning formula, surface confluence, federated learning, trusted execution environments, optimal mechanisms, multi-attribute auctions, multi-scale loss, scenario reconfiguration, probabilistic models, topology reconfiguration models, etc., in scenarios of flood prediction, social recommendation, multi-auction, terrorist attack prediction, etc. We believe that these studies are valuable in this field.