Towards a semantic layer for Italian emergency plans

Emergency plans require a complex collaboration among multiple departments and roles. They are generally long textual documents containing practical instructions for hazard responses in natural language. This work focuses on converting informal documents to a more rigorous structured-text representation by taking advantage of well-known techniques from the literature. However, this task is costly, it requires technical skills and sound domain knowledge, and it is entirely subjective. To this aim, we propose a semantic layer that supports the formalization of an emergency plan by identifying essential elements of the input document and highlighting inconsistencies, redundancies, and ambiguities.