Detection of Unstructured Sensitive Data Based on a Pre-Trained Model and Lattice Transformer

Sensitive data often contains a significant amount of unstructured information, posing great challenges to its processing. When dealing with unstructured data, effectively utilizing the structure of this sensitive data without a fixed format can greatly enhance the accuracy of sensitive data detection. Therefore, in this paper, we have developed and evaluated a Chinese sensitive data detection method for unstructured data using a pre-trained model and a Lattice-Transformer structure. This method leverages the lattice structure to fully utilize the structural information of sensitive data, overcomes the complexity and uncertainty of unstructured information through the powerful parallel computing capability of the transformer, and benefits from the strong language representation ability of the pre-trained model, enabling the method to possess strong detection capabilities for unstructured sensitive data. The outcomes of the experiment demonstrate that the model is capable of adeptly managing sensitive data in Chinese, regardless of its unstructured nature, type, or provenance.