Statistical Foundations of Learning

What it is about

You learn statistical foundations of machine learning theory and mathematical tools for the analysis of learning algorithms. In the end you will be able to evaluate generalization and consistency questions, theoretically analyze algorithms such as k‑NN, SVM and simple neural networks, and contextualize newer developments such as overparameterization and training dynamics.

What you will be able to do

• Mathematically analyze classification and clustering algorithms
• Knowledge of statistical techniques (limit theorems, concentration inequalities)
• Apply statistical tools to improve algorithms
• Assess the limits of formulations and theories in machine learning

What the module consists of

• VorlesungTheoretical foundations and current research findings
• Übung/Tutorialweekly tutorials on discussing example problems
• Assignmentsbiweekly exercises for deeper understanding; optionally with bonus effect

Teaching method

• Lecture slides with occasional blackboard workPresentation of central theoretical frameworks and proofs
• Tutorials with model solutionsApplication of theory to exercise examples and derivation of new results
• Bi‑weekly tasksindividual or pair exercises for deeper understanding; graded tasks may grant a bonus
• Moodle and live streamingasynchronous discussion, material provision and participation options