Methods of semantic annotation and ontological modeling of mathematical texts in LaTeX format

Problem statement. This work addresses the problem of structuring scientific articles for building data corpora within the SciLibAIRU semantic ecosystem, and the transition from a document-oriented data representation to a format suitable for automated analysis and semantic search. Objective. The objective of this work is to investigate an optimal method for vector-based fuzzy search over mathematical texts and to implement it in conjunction with a parser for mathematical LaTeX texts. Results. A prototype vector search system is proposed that is capable of ingesting LaTeX versions of scientific texts and providing a fuzzy search interface over textual fragments within the SciLibAIRU library. Practical significance. The results are applicable to library and editorial information systems.