Sommersemester 2025
Bachelor
Projektgruppen
Maschinelles Lernen
Bachelor BA-INF 051
Contact
Details
Preliminary Meeting
Mittwoch, 09. April 2025
10 Uhr s.t.
Institut für Informatik
Raum 3.110
Participants
max. 6
Prerequisites
keine
Registration
Bitte registrieren Sie sich für den Kurs in ecampus vor dem Preliminary Meeting, siehe den Button unten.
Wissensentdeckung
Bachelor BA-INF 051
Lecturers
Prof. Dr. Stefan Wrobel, Dr. Tamas Horvath, Fouad Alkhoury, Sebastian Müller, Vanessa Toborek
Contact
Details
Preliminary Meeting
Mittwoch, 09. April 2025
10 Uhr s.t.
Institut für Informatik
Raum 3.110
Participants
max. 6
Registration
Bitte registrieren Sie sich für den Kurs in ecampus vor dem Preliminary Meeting, siehe den Button unten.
Lernen mit Graphen
Bachelor BA-INF 051
Contact
Details
Preliminary Meeting
Mittwoch, 09. April 2025
10 Uhr s.t.
Institut für Informatik
Raum 3.110
Participants
max. 6
Registration
Bitte registrieren Sie sich für den Kurs in ecampus vor dem Preliminary Meeting, siehe den Button unten.
Vorlesungen
Grundlagen der künstlichen Intelligenz I
Bachelor BA-INF 160
Dieses Modul ist dem maschinellen Lernen gewidmet, einem der wichtigsten Bereiche der
künstlichen Intelligenz. Die Studierenden lernen, implementieren und üben die wichtigsten
Algorithmen des maschinellen Lernens. Das Modul konzentriert sich auf die Kernaufgaben des
prädiktiven Lernens aus Beispielen und des Agentenlernens und lehrt die wichtigsten Klassen
von Algorithmen für diese Aufgaben. Am Ende des Moduls sind die Studierenden in der Lage,
geeignete Methoden und Systeme für spezifische Anwendungen des prädiktiven Lernens
auszuwählen, einzusetzen und, wenn nötig, anzupassen oder weiterzuentwickeln.
Inhalte:
Verschiedene Arten von Lernproblemen, wichtige nicht-parametrische und parametrische
Methoden für überwachtes Lernen (z.B., Konzeptlernen als Suche in geordneten
Hypeothesenräumen, lernen von Entscheidungsbäumen, probabilistische Ansätze, Kernel-
Methoden, lineare und logistische Regression, gradient descent, neuronale Netze, deep
learning), Lerntheorie.
Lecturers
Contact
Details
Vorlesung - Start
10. April 2025
donnerstags, 12:15 - 13:45
Friedrich-Hirzebruch Allee 5 - Hörsaal 2
Übungen - Start
24. April 2025
donnerstags
Gruppe I: 14:15 - 15:45 Uhr, Room 1.047
Gruppe II: 14:15 - 15:45 Uhr, Room 2.113
Gruppe III: 16:15 - 17:45 Uhr, Room 1.047
Gruppe IV: 16:15 - 17:45 Uhr, Room 2.025
Tutoren
Simon Kraft, Felix Müller, Denis Schafranski
Prerequisites
keine
Registration
Bitte registrieren Sie sich für den Kurs in ecampus, siehe den Button unten.
Important Dates
tba
tba
Master
Lectures
Quantum Computing Algorithms
Master MA-INF 1224
Tba
Lecturers
Contact
Details
Lecture - Start/Time/Place
07. April 2025
Mondays, 12:15 - 13:45
Hörsaal IV, Meckenheimer Allee 176
Exercises - Start/Time/Place
02. Mai 2025
Fridays, 14:15 - 15:45
Hörsaal IV, Meckenheimer Allee 176
Registration
Please register in ecampus
Important Dates
tba
Seminars
Principles of Data Mining and Learning Algorithms: Selected Papers from the state-of-the-art
Master: MA-INF 4209
The seminar will held as a block seminar. Further seminar dates will be announced during the preliminary meeting.
Lecturers
Prof. Dr. Stefan Wrobel, Dr. Tamas Horvath, PD Dr. Michael Mock
Contact
Details
Preliminary Meeting
Wednesday, 09 April 2025
10:15 h
Institut für Informatik, Raum 1.033
Friedrich-Hirzebruch-Allee 8
Participants
max. 6
Prerequisites
none
Registration
Please register in ecampus for the course before the preliminary meeting, see the button below.
Bridging Philosophy and AI: Theories of Explanation and Explainability in Deep Learning
Master: MA-INF 4209
As machine learning models become more and more sophisticated, there is growing interest in finding ways to explain their behavior to make it less opaque and more understandable. But what is an explanation? What does it mean to explain something or to understand it? In this seminar we study philosophical theories of scientific explanation and understanding and explore how to apply them to computational approaches to explainability for deep learning and other neural network architectures. The seminar does not presume a deep background in philosophy or technical training in AI, and is suitable for all students of computer science or philosophy who would like to learn more about how philosophical ideas relate to developments in AI.
Contact
Details
Preliminary Meeting
Thursday, April 17, 2025
14:00
Institut für Informatik, Room 2.050
Friedrich-Hirzebruch-Allee 8
Participants
max. 10
Prerequisites
none
Registration
Please register before the preliminary meeting in eCampus.
Labs
Development and Application of Data Mining and Learning Systems: Machine Learning
Master: MA-INF 4306
Contact
Details
Preliminary Meeting
Wednesday, 09 April 2025
11:00 h s.t.
Room 3.110,
Friedrich-Hirzebruch-Allee 8
Participants
max. 6
Prerequisites
MA-INF 4212 highly recommended
Registration
Please register in ecampus for the course before the preliminary meeting, see the button below.
Development and Application of Data Mining and Learning Systems: Data Mining
Master: MA-INF 4306
Contact
Details
Preliminary Meeting
Wednesday, 09 April 2025
11:00 h s.t.
Room 3.110,
Friedrich-Hirzebruch-Allee 8
Participants
max. 6
Prerequisites
MA-INF 4212 highly recommended
Registration
Please register in ecampus for the course before the preliminary meeting, see the button below.
Development and Application of Data Mining and Learning Systems: On neural operators for in-context/zero-shot function estimation
Master: MA-INF 4306
Slides from Preliminary Meeting
(In Basis: Lab Development and Application of Data Mining and Learning Systems: Neural ODEs)
In-context methods estimate probability distributions (over datasets, parameters or functions) by leveraging pre-trained deep neural network models trained on some large and varied datasets.
While LLMs are the most prominent examples of in-context models, recent work has demonstrated good performance of in-context methods in more classical machine learning tasks like regression and classification.
In this lab, we will (try to) develop in-context models for function estimation (e.g. time series interpolation) and compare them to classical approaches. Concretely, we will use transformer networks and neural operators (i.e. neural networks that learn mappings between infinite dimensional spaces) and compare against Gaussian process regression models.
You will have to:
- read and discuss some of the relevant literature
- implement (code) and test these in-context models and baselines
- give presentations about your experiments and hand in a final report (working in groups, if possible)
Requirements: working knowledge of machine learning libraries like PyTorch or Jax.
Lecturers
Contact
Details
Preliminary Meeting
Tuesday, 08 April 2025
17:00 s.t.
Participants
max. 6
Prerequisites
Working knowledge of machine learning libraries like PyTorch or Jax
Registration
Slides from Preliminary Meeting
Send us an email with your full name, matriculation number and (if you want to) discord username.