The Artificial Intelligence Ontology: LLM-assisted construction of AI concept hierarchies
Project Overview
The document explores the transformative role of generative AI in education, highlighting the development and implications of the Artificial Intelligence Ontology (AIO), a comprehensive framework designed to standardize AI concepts and methodologies. Created with the help of large language models (LLMs), AIO addresses both technical and ethical dimensions of AI, effectively supporting researchers, developers, and educators by providing a structured approach to understanding AI. Its practical applications include annotating AI research publications and enhancing bioinformatics resources, which serve to foster transparency and comprehension within the AI research community. The findings indicate that AIO not only streamlines the integration of AI in educational contexts but also promotes ethical considerations and best practices, ultimately enhancing the learning experience and facilitating informed decision-making among stakeholders in the field. Through these initiatives, the document underscores the potential of generative AI to enrich education by providing innovative tools and resources that empower educators and learners alike.
Key Applications
Artificial Intelligence Ontology (AIO)
Context: AI researchers, developers, and educators seeking standardized terminology and concepts
Implementation: Developed using manual curation and LLM assistance, structured with the Ontology Development Kit (ODK), and dynamically updated for relevance
Outcomes: Enhanced understanding of AI concepts, improved communication within the research community, and facilitated cross-disciplinary research
Challenges: Complexity of AI technologies and the need for ongoing updates to keep the ontology relevant
Implementation Barriers
Technical barrier
Complexity of AI technologies makes it difficult to cover all specific implementations and parameter values. AIO does not delve into specific model implementations or parameters, which may limit its applicability in detailed AI projects.
Proposed Solutions: Focus on broad AI concepts rather than individual model specifics to manage complexity. Future developments may address detailed model representation and integration of parameters.
Project Team
Marcin P. Joachimiak
Researcher
Mark A. Miller
Researcher
J. Harry Caufield
Researcher
Ryan Ly
Researcher
Nomi L. Harris
Researcher
Andrew Tritt
Researcher
Christopher J. Mungall
Researcher
Kristofer E. Bouchard
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Marcin P. Joachimiak, Mark A. Miller, J. Harry Caufield, Ryan Ly, Nomi L. Harris, Andrew Tritt, Christopher J. Mungall, Kristofer E. Bouchard
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai