Machine Learning
Prof. Dr. Josif Grabocka, Prof. Dr. Abhinav Valada, Dr. Daniele Cattaneo
Co-Organizers:
Maximilian Luz, Maciej Janowski, Gresa Shala
This course provides you with a good theoretical understanding and practical experience about the basic concepts of machine learning. You shall be enabled to implement a number of basic algorithms, understand advantages and drawbacks of single methods and know typical application domains thereof. Furthermore, you should be able to use (Python) software libraries in order to work on novel data analysis problems. The course will prepare you to dive deeper into advanced methods of ML, e.g. deep learning, recurrent networks, reinforcement learning, hyperparameter optimization, and into specific application domains such as image analysis, brain signal analysis, robot learning, bioinformatics etc., for which specialized courses are available.
| Time: | Monday, 10:00-12:00 First meeting on October 16th 2023. |
| Location: | This course will be taught in person. Weekly teaching will be held on Monday at HS 00 026 (G.-Köhler-Allee 101). Exercise sessions will take place on Friday 14:00-16:00 at HS 00 006 (G.-Köhler-Allee 082) |
| Learning Platform: | ILIAS |
Course Overview
The course will be taught in English
Every week there will be:
- an in-person lecture (Mondays, 10:00-12:00)
- an exercise sheet
- an in-person exercise session (Fridays 14:00 - 16:00)
At the end, there will be a written exam.
Exercises should be completed in groups and submitted a week (+ 2 day) after their release. Submissions will not be graded but they are crucial for a deeper understanding of the topics taught in the lecture. Every Friday during the exercise session, the solution of the previous' week assignment will be discussed.
We additionally provide video recordings of old lectures online, but we strongly encourage students to attend the in-person lectures (HS 00 026, G.-Köhler-Allee 101) and exercise sessions (HS 00-006, G.-Köhler-Allee 082), as the lectures' content might differ from the one in the videos.
Exam: The exam will be a closed-book written exam. Past exams can be found at ILIAS.
Course Schedule
The following are the dates for the in-person lectures:
16.10.23 – Lecture 1: Introduction
23.10.23 – Lecture 2: Linear Methods
30.10.23 – Lecture 3: Principles of Regularization
06.11.23 – Lecture 4: Support Vector Machines
13.11.23 – Lecture 5: Decision Trees
20.11.23 – Lecture 6: Neural Networks (I)
27.11.23 – Lecture 7: Neural Networks (II)
04.12.23 – Lecture 8: Ensembles
11.12.23 – Lecture 9: Gradient-Boosted Trees
18.12.23 – Lecture 10: Hyper-parameter Tuning
08.01.24 – Lecture 11: Fighting Overfitting
15.01.24 – Lecture 12: Error Metrics
22.01.24 – Lecture 13: Self- and Un-supervised Techniques
29.01.24 – Lecture 14: Round-up / Exam Q & A
The course material (old lecture video, slides, exercise sheet) for “Lecture N” will be made available the same day as the in-person lecture. For example, the material for Lecture 2 will be available on 23.10.23, and solutions to the exercises must be submitted latest 01.11.23 at 23:59.
In the first session (on 16.10.23) you will get additional information about the course and get the opportunity to ask general questions (and form groups!) While there is no need to prepare for this first session, we encourage you to already think about forming teams.
Questions?
If you have a question, please post it in the ILIAS forum (so everyone can benefit from the answer).