Administrative Information
| Title | Introduction to General Explainable AI |
| Duration | 60 mins |
| Module | B |
| Lesson Type | Lecture |
| Focus | Ethical - Trustworthy AI |
| Topic | General Explainable AI |
Keywords
Explainable Artificial Intelligence,Machine Learning,Deep Learning,Interpretability,Comprehensibility,Transparency,Privacy,Fairness,Accountability,Responsible Artificial Intelligence,
Learning Goals
- Understand, analyze and elaborate upon the importance of XAI in the modern world.
- Differentiate between transparent and opaque machine learning models.
- Categorize and discuss approaches to explainability XAI based on model scope, agnosticity, data types and explanation techniques.
- Discern, investigate and discuss the trade-off between accuracy and interpretability.
- Summarize and understand the working principles and mathematical modeling of XAI techniques like LIME, SHAP, DiCE, LRP, counterfactual and contrastive explanations.
- Expand on possible applications of XAI techniques like LIME, SHAP, DiCE, LRP to generate explanations for black-box models for tabular, textual, and image datasets.
Expected Preparation
Learning Events to be Completed Before
None.
Obligatory for Students
- Fundamentals of Python Programming
- Fundamentals of Machine Learning
Optional for Students
- Explainable Artificial Intelligence: An Introduction to Interpretable Machine Learning (1st ed. 2021 Edition) by Uday Kamath (Author), John Liu (Author)
References and background for students
Recommended for Teachers
- Belle V., Papantonis I., Principles and Practice of Explainable Machine Learning
- Arrieta, A. B. et al., Explainable Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities and Challenges Toward Responsible AI
- Explainable Artificial Intelligence: An Introduction to Interpretable Machine Learning (1st ed. 2021 Edition) by Uday Kamath (Author), John Liu (Author)
Lesson materials
Instructions for Teachers
This lecture provides general insights into the field of Explainable Artificial Intelligence (XAI). Our reliance on artificial intelligence models is further discussed. Teachers can emphasize that recent laws have also caused the urgency about explaining and defending the decisions made by AI systems. This lecture discusses tools and techniques to visualize, explain, and build trustworthy AI systems.
Outline
| Duration | Topic | Description |
|---|---|---|
| 5 mins | Introduction | Definition of XAI. Why is XAI Important and What Problems Does It Solve. |
| 5 mins | Dimensions of Explainability | What Does Explainability Mean. What Criteria Does It Have to Answer. |
| 20 mins | Approaches to Explainability | Transparent Models and Opaque Models. |
| 20 mins | Explainability Techniques | Approaching Explainability with Model-Specific and Model-Agnostic Techniques. |
| 10 mins | Closing Remarks | Discussion with Students. Questions and Answers. |
Acknowledgements
This presentation was developed by Christina Todorova and Dr George Sharkov at the European Software Institute – Center Eastern Europe.
The presentation is heavily based on and uses materials and structure from the work of Belle V., Papantonis I., "Principles and Practice of Explainable Machine Learning“ and Arrieta, A. B. et al., “Explainable Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities and Challenges Toward Responsible AI”. Please, consider reading their original research.
The Human-Centered AI Masters programme was Co-Financed by the Connecting Europe Facility of the European Union Under Grant №CEF-TC-2020-1 Digital Skills 2020-EU-IA-0068.