A FAIR and AI-ready Higgs boson decay dataset

Area of Science:

• High Energy Particle Physics
• Data Science
• Artificial Intelligence

Background:

• Massive scientific datasets require adherence to FAIR principles for effective reuse by humans and machines.
• Existing methods for assessing data and AI model FAIRness are often domain-specific.
• The FAIR principles (Findability, Accessibility, Interoperability, Reusability) are crucial for modern scientific research.

Purpose of the Study:

• To provide a domain-agnostic, step-by-step guide for evaluating the FAIRness of datasets and AI models.
• To demonstrate the application of this guide using a real-world dataset from the CERN Large Hadron Collider.
• To facilitate the creation of FAIR AI models and datasets in high energy particle physics.

Main Methods:

• Development of a domain-agnostic assessment guide for FAIR principles.
• Application of the guide to a simulated dataset of Higgs boson decays and background from the CMS Collaboration.
• Utilized available tools and community feedback for validation of the FAIRness assessment.

Main Results:

• A comprehensive guide to assess the FAIRness of scientific data and AI models was developed.
• The guide was successfully applied to a CERN open dataset, demonstrating its practical utility.
• The study provides a reproducible workflow for evaluating FAIR data principles.

Conclusions:

• The developed guide effectively evaluates the FAIRness of scientific datasets.
• This work supports the broader goal of making scientific data and AI models more reusable.
• This is the first in a series of publications aimed at promoting FAIR AI and data practices in particle physics.