Semantic Finder

Semantic Finder enables engineers to efficiently find the information they need in technical documentation, regardless of specific search terms (searching by the meaning of a phrase). This is achieved through full-text and semantic search (powered by Transformer language models) across various types of indexed documents or other text data sources. Users can obtain answers to their questions based on the search results (Question Answering). During the project, we conducted an evaluation of the search methods and language models involved with engineers using technical document datasets with which they are familiar. The language models evaluated include pre-trained Transformer models, Transformer models fine-tuned on domain-specific documents, Sentence Transformer models, and multilingual models.