Vorlesung: Embodied Artificial Intelligence - Details

Vorlesung: Embodied Artificial Intelligence - Details

Nachrichtenbox schließen
Sie sind nicht in Stud.IP angemeldet.

Allgemeine Informationen

Veranstaltungsname Vorlesung: Embodied Artificial Intelligence
Veranstaltungsnummer INF-0509
Semester SS 2025
Aktuelle Anzahl der Teilnehmenden 4
erwartete Teilnehmendenanzahl 18
Heimat-Einrichtung Professur für intelligente Perzeption in technischen Systemen - Prof. Dr. Jörg Stückler
Veranstaltungstyp Vorlesung in der Kategorie Lehre
Erster Termin Mittwoch, 23.04.2025 12:15 - 13:45, Ort: (1054 N)
Veranstaltung findet in Präsenz statt / hat Präsenz-Bestandteile Ja
Hauptunterrichtssprache englisch
ECTS-Punkte 8

Räume und Zeiten

(1054 N)
Mittwoch: 12:15 - 13:45, wöchentlich (14x)
(F1 202)
Montag, 01.09.2025 - Freitag, 05.09.2025, Montag, 08.09.2025 - Freitag, 12.09.2025, Montag, 15.09.2025 - Freitag, 19.09.2025, Montag, 22.09.2025 - Freitag, 26.09.2025, Montag, 29.09.2025 09:00 - 18:00

Modulzuordnungen

Kommentar/Beschreibung

Robotics research is currently making significant progress due to advances in learning methods, foundation models, highly parallizable simulations, and capable robot hardware. This opens up new perspectives for robotics applications in everyday environment such as our homes or in flexible production.

These developments are often summarized as "embodied artificial intelligence". "Embodied" means that robot intelligence strongly depends on the capabilities of the robot in terms of sensors, actuators, and body shape, and that intelligence arises by learning in a sensing-and-action loop. This is different to classical AI methods, which abstract away the physical world (for example, symbolic blocks world) and typically break in unforeseen situations.

This course teaches fundamentals and contemporary methods for embodied artificial intelligence. Topics include modern reinforcement learning, imitation learning, foundation models for robotics, and scene perception. It consists of a lecture part and an exercise part in which students will gain practical experience in team projects on robot learning and perception methods.

While the lecture will be given during the lecture period, the team project exercises will be conducted in a block in the lecture free period. The first two weeks of the team projects will be densely tutored and will have a fixed time schedule. The remaining time (about two weeks) until the team project presentation will be sparsely tutored and the time schedule for working on site in the lab is more flexible. Detailed information on the course schedule and organization will be provided in the preliminary meetings (first and second lecture) in the first lecture week.

The number of participants in this course is limited. Please register preliminarily in the course if you are interested to participate. Places in the course will be assigned by the course organizers after the lecture in the second (note: updated!) lecture week. We will introduce the organizational details of the course in the first and second lecture slot.

Prerequisites for this course:
- Basic programming knowledge in Python
- Basic knowledge in Deep Learning
- Recommended: Successful participation in modules INF-0508 “Probabilistic Machine Learning” or INF-0476 “Computer Vision for Intelligent Systems”

Anmeldemodus

Die Auswahl der Teilnehmenden wird nach der Eintragung manuell vorgenommen.

Please register preliminarily if you are interested to participate in this course.

The number of participants in this course is limited. Places in the course will be assigned by the course organizers after the lecture in the second (note: updated!) lecture week. Participation in one of the preliminary meetings (first or second lecture) is mandatory.