Tutorials for eXplainable Artificial Intelligence (XAI) methods

“The problem is that a single metric, such as classification accuracy, is an incomplete description of most real-world tasks.” — (Doshi-Velez et al., 2017)

This repository contains a collection of interactive tutorials for understanding and applying modern eXplainable Artificial Intelligence (XAI) methods to machine learning and deep learning models, including Random Forests, CNNs, and Transformers. Each tutorial is provided as a Jupyter Notebook combining short video lectures with practical hands-on exercises.

The tutorials cover both model-agnostic and model-specific XAI methods, including SHAP, LIME, Permutation Feature Importance, Grad-CAM, Attention Maps, and Forest-Guided Clustering.

The learning objectives are:

• understand the importance of interpretability and transparency in AI

• learn how different XAI methods work and when to use them

• interpret explanation outputs and visualizations for different model types

• gain hands-on experience applying XAI methods to real-world examples

📚 Theoretical Background

EXPLAINABLE AI

• Terminology
• Why is explainability important?
• Taxonomy of XAI methods

MODEL-AGNOSTIC XAI METHODS

• Introduction to Permutation Feature Importance
• Introduction to Local Interpretable Model-Agnostic Explanations (LIME)
• Introduction to SHapley Additive exPlanations (SHAP)

MODEL-SPECIFIC XAI METHODS

• Introduction to Forest-Guided Clustering
• Introduction to Grad-CAM
• Introduction to Attention Maps

MACHINE LEARNING BASICS

• Introduction to Random Forest Models
• Introduction to Convolutional Neural Networks and Feature Visualization
• Introduction to Transformers

🤝 Contributing

Comments, suggestions, and contributions are very welcome! If you have ideas for improvements or want to report an issue, feel free to open an issue or submit a pull request. We will review your contribution as soon as possible.

🛡️ License

This repository contains both source code and teaching materials, which are licensed separately:

• Code (notebooks, scripts, and other software) is licensed under the MIT License. See the LICENSE file.

• Teaching materials (slides, figures, and written explanations) are licensed under the Creative Commons Attribution 4.0 License. See the LICENSE-CONTENT file.

https://creativecommons.org/licenses/by/4.0/