Data Analysis Block Course: Statistics and Machine learning

Name: Data Analysis Block Course: Statistics and Machine learning
Start: 2026-09-21T09:00:00+02:00
End: 2026-10-02T20:00:00+02:00
Location: TUD/RUB/Hybrid

21 Sept 2026, 09:00 → 2 Oct 2026, 20:00 Europe/Zurich

TUD/RUB/Hybrid

TUD: Otto-Hahn-Straße 4a, 44227 Dortmund, CP-03-123 RUB: Universitätstr. 150, 44780 Bochum, NB2/158

Description

Moodle page: Bochum, Dortmund
ecampus: Bochum (160435),

Literature:

G. Cowan , Statistical Data Analysis
O. Behnke et al., Data Analysis in High Energy Physics

Previous courses:

Previous course: 2025 StatML block course
RUB course, (git repository)

Monday 21 September
- 09:00 → 10:30
  
  Lecture - Introduction 1h 30m
  
  Speaker: Michael Schmelling (Max Planck Society (DE))
- 10:45 → 12:00
  
  Exercises 1h 15m
  
  Speaker: Jan Peter Wagner (Technische Universitaet Dortmund (DE))
- 13:30 → 14:30
  
  Lecture - Error propagation 1h
  
  Speaker: Michael Schmelling (Max Planck Society (DE))
- 14:45 → 16:00
  
  Exercises 1h 15m
  
  Speaker: Jan Peter Wagner (Technische Universitaet Dortmund (DE))
Tuesday 22 September
- 09:00 → 10:30
  
  Lecture - Information, inference and confidence level intervals 1h 30m
  
  Speaker: Michael Schmelling (Max Planck Society (DE))
- 10:45 → 12:00
  
  Exercises 1h 15m
  
  Speaker: Jan Peter Wagner (Technische Universitaet Dortmund (DE))
- 13:00 → 14:00
  
  Lecture - Least Squares Fitting 1h
  
  Speaker: Michael Schmelling (Max Planck Society (DE))
- 14:15 → 15:30
  
  Exercises 1h 15m
  
  Speaker: Jan Peter Wagner (Technische Universitaet Dortmund (DE))
Wednesday 23 September
- 09:00 → 10:30
  
  Lecture - Maximum Likelihood, Sandwich estimator, Bootstrap 1h 30m
  
  Speaker: Michael Schmelling (Max Planck Society (DE))
- 10:45 → 12:00
  
  Exercises 1h 15m
  
  Speaker: Jan Peter Wagner (Technische Universitaet Dortmund (DE))
- 13:00 → 14:00
  
  Lecture - Markov Chain Monte Carlo, COWs 1h
  
  Speaker: Michael Schmelling (Max Planck Society (DE))
- 14:15 → 15:30
  
  Exercises 1h 15m
  
  Speaker: Jan Peter Wagner (Technische Universitaet Dortmund (DE))
Thursday 24 September
- 09:00 → 10:30
  
  Lecture - sPlot, sFit, Unfolding 1h 30m
  
  Speaker: Michael Schmelling (Max Planck Society (DE))
- 10:45 → 12:00
  
  Exercises 1h 15m
  
  Speaker: Jan Peter Wagner (Technische Universitaet Dortmund (DE))
- 13:00 → 14:00
  
  Lecture - Unfolding, Harmonic analysis 1h
  
  Speaker: Michael Schmelling (Max Planck Society (DE))
- 14:15 → 15:30
  
  Exercises 1h 15m
  
  Speaker: Jan Peter Wagner (Technische Universitaet Dortmund (DE))
Friday 25 September
- 09:00 → 10:30
  
  Automatic Differentiation 1h 30m
  
  Speaker: Prof. Mikhail Mikhasenko (Ruhr Univeristy Bochum)
- 10:45 → 12:15
  
  Exercises - Automatic Differentiation 1h 30m
  
  Exercise notebook
  
  Solution notebook
- 13:30 → 15:00
  Machine Learning (loss, ROC) 1h 30m
  - Slides
  - lecture-1B-entropy.jl: introduction to entropy
  - lecture-1B-mse.jl: MSE as the simplest loss function
  Speaker: Prof. Mikhail Mikhasenko (Ruhr Univeristy Bochum)
- 15:15 → 16:45
  
  Exercises - Samples, FoM, ROC 1h 30m
  
  Exercise notebook
  
  Solution notebook
Monday 28 September
- 09:00 → 10:30
  Decision Trees, Adaboost 1h 30m
  - Slides
  - lecture-2A-split.jl: splitting index demonstration
  - lecture-2A-trees.jl: decision tree demonstration
  Speaker: Prof. Mikhail Mikhasenko (Ruhr Univeristy Bochum)
- 10:45 → 12:15
  
  Exercises - Decision Trees 1h 30m
  
  Exercise notebook
  
  Solution notebook
- 13:30 → 15:00
  Decision Trees, GradientBoost 1h 30m
  - Slides
  - GBDT playground
  Speaker: Prof. Mikhail Mikhasenko (Ruhr Univeristy Bochum)
- 15:15 → 16:45
  
  Exercises - AdaBoost, GradientBoost 1h 30m
  
  Exercise notebook
  
  Solution notebook
Tuesday 29 September
- 09:00 → 10:30
  Neural Networks : General introduction 1h 30m
  - Slides
  - lecture-3A-xor.jl: learning xor demonstrator
  - gradient descent app: github.com/lilipads/gradient_descent_viz
  Speaker: Prof. Mikhail Mikhasenko (Ruhr Univeristy Bochum)
- 10:45 → 12:15
  
  Exercises - Neural Networks 1h 30m
  
  Exercise notebook
  
  Solution notebook
- 13:30 → 15:00
  Neural networks : Bias-Variance TradeOff and regularization techniques 1h 30m
  - Slides: Lecture 3 pdf contains but A,B parts
  - lecture-3B-UAT.jl: Universal approximation theory demonstration
  - lecture-3B-NMIST.jl: training of MNIST classification
  - playground.tensorflow.org
  Speaker: Prof. Mikhail Mikhasenko (Ruhr Univeristy Bochum)
- 15:15 → 16:45
  
  Exercises - NNs on example data 1h 30m
  
  Exercise notebook
  
  Solution notebook
Wednesday 30 September
- 09:00 → 11:00
  Generative networks, Normalizing flow 2h
  - Slides
  - lecture-4A-transformations.jl: demonstration of bijectors
  - lecture-4A-RQS.jl: tuning RQS
  Speaker: Prof. Mikhail Mikhasenko (Ruhr Univeristy Bochum)
- 11:15 → 12:30
  
  Exercises - Normalizing flow 1h 15m
  
  Notebook NF basics
  
  Notebook NF training
- 16:30 → 19:35
  
  BBQ 3h 5m
  
  Sign up here: https://hackmd.e5.physik.tu-dortmund.de/3XzYG8C4ScuKaNh43QB3GQ?both
  
  Location: https://maps.app.goo.gl/qwZvu8eUB14JzKTw5
Friday 2 October
- 09:00 → 10:30
  Written Exam -- date to be fixed 1h 30m
  - Pen and paper, calculator (non-programmable)
  - Time: 1:30h
  - 5-6 Problems
  - ChatGPT is not allowed