The method for adaptive automatic selection of text content search modes based on entity distribution in generative artifical intelligence systems

Abstract

The method for adaptive automatic selection of text content search modes based on entity distribution in generative artificial intelligence systems has as a novelty a phase for preparing text data with: step (100) for receiving text documents, step (101) for segmenting text content into segments of smaller length, step (102) for converting segments into vector representations using a language model based on deep learning, step (103) for storing the obtained vectors, step (104) where triples in the form of subject-relation-object are generated using a large language model for automatic recognition of entities and relations between them, step (105) for forming a map that assigns to each unique entity a list of identifiers of segments, and step (106) for constructing a knowledge graph based on separated triplets. Method also includes a phase for processing a user query with: step (200) for receiving a user query, step (201) for identifying an entity in the user query, step (202) for obtaining segment identifiers based on the entity using the map from step (105), step (203) for automatically selecting a text content search mode for generating a response.