Data privacy and audit compliance of supply chain finance warehouse receipt pledge based on WSN federated edge learning
Year: 2026
Type: preprint
Abstract: In supply chain finance warehouse receipt pledge financing, the status of the pledged assets is difficult to verify in real time, the risk of privacy leakage of financing data is high, and the federal training process lacks verifiable audit trajectories for regulatory authorities. To achieve multi-warehouse collaborative risk control modeling under the condition of data not leaving the domain and meet the requirements of privacy protection and audit compliance, a method for privacy and audit compliance of supply chain finance warehouse receipt pledge data based on federated edge learning of wireless sensor networks is constructed. A pledge perception and digital warehouse receipt generation mechanism based on wireless sensor networks is designed. A federated edge learning architecture integrating trust and energy-aware weighted aggregation, federated proximal terms and communication compression is constructed. The study introduces a joint security aggregation mechanism of differential privacy and homomorphic encryption, as well as an audit compliance framework for blockchain smart contracts. The experimental results show that when the data heterogeneity parameter is 0.1, the test accuracy of the proposed method reaches 88.4%, which is 9.9% higher than the 78.5% of the federal average algorithm. In the ablation experiment, the F1 score of the complete method is 93.6%, while it drops to 88.9% after removing the federal proximal term. When the privacy budget is 0.10, the accuracy rate of the proposed method is 92.1%, which is 8.0% higher than the 84.1% of the federal average algorithm. The results show that the proposed method has stronger robustness under extreme non-independent homogeneous distribution conditions. Trust and energy-aware selection and federated proximal terms contribute significantly to accuracy, and the privacy enhancement mechanism achieves effective privacy protection with limited accuracy loss. This research provides a technical path that takes into account data privacy, model performance and audit compliance for the warehouse receipt pledge scenario in supply chain finance, which is conducive to promoting the compliant application of federated learning in supply chain finance risk control.
Source: Research Square
Institution: Shandong University of Technology
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7225289911