LinkedIn
Facebook
X (Twitter)
WhatsApp
Beschreibung
A comprehensive, hands-on course reference and lab environment for Explainable Artificial Intelligence (XAI), Machine Learning Model Interpretability, and Responsible AI.
Learning Goals:
- Distinguishing interpretable models from post-hoc explanations of black-box models.
- Explaining the mathematics behind Shapley values, LIME and gradient-based attribution.
- Applying SHAP, LIME and Captum to tabular, image and text data.
- Interpreting model decisions to uncover key features, biases and errors.
- Auditing a model of their own with XAI methods and presenting the findings.
Course Overview
This course provides a systematic, practical introduction to the explainability and interpretability of complex machine learning and deep learning models. It addresses the "black box" model problem and equips students with mathematical and applied methods to make model decisions transparent, understandable, and auditable.
The accompanying GitHub repository serves as a central reference and workshop environment. Through interactive Jupyter Notebooks, participants explore theoretical concepts directly in Python and apply established XAI libraries (such as SHAP, LIME, and Captum) to tabular, image, and text datasets.
Prerequisites
To successfully participate in the course and work through the notebooks, the following background knowledge is required:
Mathematics: Basic knowledge of Statistics, Probability Theory, and Linear Algebra (vectors, matrices, partial derivatives). Programming: Proficiency in Python (NumPy, Pandas, Matplotlib, Seaborn) and experience with Jupyter Notebooks. Machine Learning: Solid understanding of classical ML (Scikit-Learn) and a basic grasp of Deep Learning (PyTorch or TensorFlow/Keras).
Termine
Montag, 26.10.2026
17:15 - 18:45
Montag, 02.11.2026
17:15 - 18:45
Montag, 09.11.2026
17:15 - 18:45
Montag, 16.11.2026
17:15 - 18:45
Montag, 23.11.2026
17:15 - 18:45
Montag, 30.11.2026
17:15 - 18:45
Montag, 07.12.2026
17:15 - 18:45
Montag, 14.12.2026
17:15 - 18:45
Montag, 11.01.2027
17:15 - 18:45
Montag, 18.01.2027
17:15 - 18:45
Online
Buchungszeitraum:
01.09.2026
-
23.10.2026