ML4DS lecture @FUB during WiSe21/22
The course provides an overview of machine learning methods and algorithms for different learning tasks, namely supervised, unsupervised and reinforcement learning. In the first part of the course, for each task the main algorithms and techniques will be covered including experimentation and evaluation aspects. In the second part of the course, we will focus on specific learning challenges including high-dimensionality, non-stationarity, label-scarcity and class-imbalance. By the end of the course, you will have learned how to build machine learning models for different problems, how to properly evaluate their performance and how to tackle specific learning challenges.
- Supervised learning
- Unsupervised learning
- Reinforcement learning
- Outlier detection
- Machine learning for high-dimensional data
- Machine learning in non-stationary environments
- Machine learning under label scarcity
- Lecturer: Prof. Dr. Eirini Ntoutsi
- TA: [M.Sc. Manuel Heurich] (http://www.mi.fu-berlin.de/inf/groups/ag-KIML/members/Scientific-Staff/Heurich/index.html)
| # | Lecture | Slides |
|---|---|---|
| 1 | Introduction | slides |
| 2 | Getting to know your data | slides |
| 3 & 4 | Supervised learning (Classification): Introduction & Decision Trees & KNNs | slides |
| 5 | Supervised learning (Classification): Naive Bayes classifiers | slides |
| 6 | Supervised learning (Classification): Evaluation | slides |
| 7 | Supervised learning (Classification): SVMs | slides |
| 8 | Supervised learning (Classification): Perceptron | slides |
| 9 | Supervised learning (Regression) | slides |
| 10 | Unsupervised learning (Clustering): Partinioning-based methods | slides |
| 11 | Unsupervised learning (Clustering): Hierarchical methods | slides |
| 12 | Unsupervised learning (Clustering): Density-based methods | slides |
| 13 | Unsupervised learning (Clustering): Evaluation | slides |
| 14 | Unsupervised learning (Clustering): EM | slides |
| 15 | Unsupervised learning (Clustering): EM | slides |
| 16 & 17 | Reinforcement learning: Introduction & MDPs | slides |
| 18 | Reinforcement learning: Model free-learning | slides |
| 19 | Reinforcement learning: Approximate Q-learning | slides |
| 20 | High Dimensionality: Feature selection | slides |
| 21 | High Dimensionality: Dimensionality reduction | slides |
| 22 & 23 & 24 | Velocity: Stream Classification | slides |
| 25 & 6 | Velocity: Stream Clustering | slides |
I tried my best to cite my sources; please let me know if you think something is missing.
Part of the material was developed/enriched for the KDD I lecture and KDD II lecture I was offering at the Department of Informatics, LMU Munich during SoSe12, WiSe15/16, respectively. I reworked and extended the material for the Data Mining I and Data Mining II lectures, I was offering at the Faculty of Electrical Engineering and Computer Science at the Leibniz University Hannover in the period 2006-2021. The RL part is an extenstion of the corresponding part of the AI lecture I was offering at the Department of Mathematics and Computer Science at the Free University of Berlin during SoSe21.