Natural Language Processing-Driven Microscopy Ontology Development
摘要
This manuscript describes the accelerated development of an ontology for microscopy in materials science and engineering, leveraging natural language processing (NLP) techniques. Drawing from a comprehensive corpus comprising over 14 k contributions to the Microscopy and Microanalysis conference series, we employed two neural network-based algorithms for NLP. The goal was to semiautomatically create the Microscopy Ontology (MO) that encapsulates and interconnects the terminology most frequently used by the community. The MO, characterized by its interlinked entities and relationships, is designed to enhance the quality of user query results within NexusLIMS. This enhancement is facilitated through the concurrent querying of related terms and the seamless integration of logical connections.