Ontologies are essential tools for structuring and collecting human knowledge. They enable the identification and description of concepts concerning a specific topic.
In the field of scientific research, ontologies enable efficient study and exploration of the global research scene and structured extraction of information from scientific articles.
Additionally, they enable the use of recognized and universally accepted terms, ensuring consistent communication and description of research.

Creating a high-quality ontology requires the analysis of several experts, who must have in-depth knowledge of the field and be able to describe the concepts and their relationships.
Furthermore, it's crucial to note that ontologies evolve over time due to the annual coinage of new scientific terms and the abandonment of others. Therefore, keeping an ontology up-to-date requires a constant commitment of these means. Currently, this process is usually done manually, taking up a lot of resources in terms of time and workforce required. The creation of an automatic method will facilitate the creation of up-to-date ontologies.

This seminar will illustrate a work-in-progress pipeline for automatically creating an ontology using state-of-the-art technologies, such as LLM models. The methodology allows the automatic creation of the list of terms the ontology accepts and classifies them to represent their relationships, effectively building the ontology. The experiments to validate the methodology rely on analyzing millions of papers in the field of Computer Science. Information extracted from these papers was used to evaluate different phases of the pipeline, revealing promising results.