Machine Learning 1 (Winter Term 2024/2025)

Overview

Course (2/2/0) consisting of:
- Lectures in TRE/PHYS/E (Zellescher Weg 16) on Fridays, 09:20–10:50
- Exercise groups
  - VMB/0302/U, Tuesdays, 16:40–18:10
  - APB/E001/U, Thursdays, 16:40–18:10
  - APB/E001/U, Fridays, 14:50–16:20
  - APB/E001/U, Fridays, 16:40–18:10
  - Online, Wednesdays, 9:20–10:50
- Self-study
- Final Examination
Lecturer: Bjoern Andres
Teaching Assistants: Jannik Irmai, David Stein
Enrolment (OPAL). Additional rules for enrolment may apply, depending on the study programme.
Forum

Contents

Introduction
Supervised learning
- Introduction
- Learning of binary decision trees
  - NP-hardness
  - Local search algorithm (exercises only)
- Learning of disjunctive normal forms (exercises only)
  - NP-hardness
- Learning of linear functions
  - Logistic regression
  - Convexity
  - Gradient descent algorithm
- Learning of composite functions (deep learning)
  - Compute graphs
  - Non-convexity
  - Forward-propagation algorithm and backward-propagation algorithm
  - Gradient descent algorithm
Semi-supervised and unsupervised learning
- Introduction
- Partitioning
  - Clique partition problem
  - NP-hardness
  - Local search algorithms
  - Relaxations and lower bounds
- Clustering
  - Multicut problem
  - NP-hardness
  - Local search algorithms
  - Relaxations and lower bounds
- Ordering
  - Linear ordering problem
  - NP-hardness
  - Local search algorithms
  - Relaxations and lower bounds
- Classifying
Supervised structured learning
- Introduction
- Conditional graphical models
- Pseudo-Boolean optimization
- NP-hardness
- Trees (a special case)
Embedding

Textbooks