Concentration inequalities provide quantitative insight how functions of (independent) random variables deviate from their expectation. In the course, we will discuss classical concentration inequalities, such as Hoeffding’s or McDiarmid’s, as well as recent extensions. In additional to the mathematical treatment, we will discuss application in machine learning and potentially other fields.
References:
Concentration Inequalities by S. Boucheron, G. Lugosi, P. Massart
Probability in High Dimension by R. Van Handel
High-Dimensional Probability by R. Vershynin
Concentration of measure by M. Ledoux
Target group: PhD students from all years, postdocs, anyone else who is interested
Prerequisites: strong background in probability and analysis
Evaluation: pass/fail based on homework
Teaching format: None
ECTS: 3 Year: 2020
Track segment(s):
CS-AI Computer Science - Artificial Intelligence
DSSC-PROB Data Science and Scientific Computing - Probabilistic Models
MAT-PROB Mathematics - Probability
Teacher(s):
Christoph Lampert Jan Maas
Teaching assistant(s):
If you want to enroll to this course, please click: REGISTER
- Trainer/in: Christoph Lampert
- Trainer/in: Jan Maas
- Teaching Assistant: Haonan Zhang