Explainable Artificial Intelligence
The target group for the course are those interested with basic to advanced knowledge in AI and machine learning. The course imparts established and current methods of explainable artificial intelligence (XAI) to improve the traceability of AI systems and make black-box models more comprehensible. Concrete use cases are used to present procedures for the explanation and evaluation of AI models and address their contribution to the development of trustworthy AI systems.
Overview
AI systems, especially those based on trained models, are often opaque black boxes. Explainable Artificial Intelligence (XAI) develops methods that make it possible to reduce this lack of transparency and make the outputs of AI systems understandable. The course methodically presents both established and current XAI approaches and illustrates them using concrete application cases. It demonstrates how the quality of XAI methods can be evaluated. In addition, it introduces the contributions XAI can make to designing trustworthy AI systems.
What content can I expect?
- Overview of approaches to explainable artificial intelligence
- Detailed introduction to fundamental XAI algorithms
- Exploring XAI methods in Jupyter Notebooks
- Trustworthiness of AI models and evaluation of XAI methods
What will I achieve?
By the end of the course, you will be able to...
- Describe various methods of explainable AI and distinguish them from each other
- Select suitable explainability methods for specific use cases
- Implement explainability methods in a use case
What prerequisites do I need?
- Basic knowledge of machine learning or artificial intelligence (recommended)
- Interest in methods of explainable artificial intelligence (XAI)
- For advanced content: basic knowledge of Python and Jupyter Notebooks is an advantage
- Basic mathematical knowledge helpful (e.g. statistics and linear algebra)
