A coal mine safety hidden danger management and control method, device, equipment and medium

Abstract

The present application discloses a kind of coal mine safety hidden danger management and control method, device, equipment and medium, it is related to coal mine safety guarantee technical field, the data set is formed after the entity and relationship annotation of text material is carried out to the coal mine safety ontology constructed by the present application, to fine-tune big language model and obtain knowledge extraction big language model, this process utilizes the text material of all directions in coal mine safety field, fine-tunes model, enhances the extraction ability of model to coal mine related knowledge;Then the semantic association between vocabulary texts is captured by calculating the attention weight between different vocabulary texts, this process is through fine-tuned enhanced knowledge extraction big language model, deeply understands the complex semantics of coal mine operation scene and captures the coal mine operation process of context association between texts, finally based on compliance review requirements, to review the operation process that does not meet safety regulations in the captured coal mine operation process, to carry out context association hidden danger analysis in complex coal mine environment.

Images