Welcome to FAIRXAI’s documentation

FAIRXAI is a platform developed within the Future Artificial Intelligence in Research (FAIR) initiative, designed to support the composition, execution, and explanation of modular AI decision-making processes.

Unlike traditional explainable AI tools that focus on individual models, FAIRXAI enables researchers and developers to describe and visualize the entire decision pipeline as a structured composition of interoperable modules. Each module contributes to the overall reasoning process and is associated with its own explanation method.

The platform empowers users to build transparent, traceable AI workflows that are aligned with the needs of scientific research, regulatory clarity, and human interpretability. FAIRXAI is the toolbox for making the next generation of AI systems not only powerful — but understandable.

• GitHub: https://github.com/kdd-lab/FAIRXAI

• Issues: https://github.com/kdd-lab/FAIRXAI/issues

Contents

• Guide to integrating an explanation method in FAIRXAI
  • 1. Architecture Overview
  • 2. Implementing the Adapter
  • 3. Handling Results (Explanation)
  • 4. Registration and Verification
  • 5. Usage via Configuration
  • Developer Checklist
• FAIRXAI API Reference
  • fairxai package
  • fairxai.app package
  • fairxai.app.explainability package
  • fairxai.bbox package
  • fairxai.data package
  • fairxai.data.dataset package
  • fairxai.data.descriptor package
  • fairxai.explain package
  • fairxai.explain.adapter package
  • fairxai.explain.explaination package
  • fairxai.explain.explainer_manager package
  • fairxai.project package
• fairxai
  • fairxai package

Indices and tables

• Index

• Module Index

• Search Page