Few shot learning characterizes a set of learning algorithms that can learn only from a few trials. These algorithms include meta-learning algorithms that can transfer knowledge to a new task, Bayesian learning methods, Bayesian optimization and simulation to reality transfer methods.

The students can choose from different topics from the area of few shot learning for robotics. Each topic consists of several research papers for which the students have to prepare a presentation as well as a report in form of a scientific research paper.

Erfolgskontrolle:

Vortrag zum gewählten Thema am Ende des Semesters und schriftliche Ausarbeitung

Arbeitsaufwand = 90 h = 3 ECTS

Präsenzzeit: 15h

Selbststudium: 45h

Scientific Report schreiben: 20h

Präsentation vorbereiten: 10h

Der Besuch der Vorlesung „Maschinelles Lernen  – Grundverfahren“ ist empfehlenswert.

Ein Rücktritt ist innerhalb von zwei Wochen nach Vergabe des Themas möglich.

It is only possible to resign within two weeks after assignment of the topic