Main content
Top content
Faculty Members
Masterseminar Incremental Learning
Informationen
- Veranstaltungsnummer:
- 6.692
- Veranstaltungsart:
- Seminar
- StudIP:
- https://studip.uni-osnabrueck.de/dispatch.php/course/details/?again=1&sem_id=f4a8bbc94fc51acf34cd1e477d530d71
- Semester:
- SoSe 2012
- ECTS-Punkte:
- 3,00
- Voraussetzungen:
Beschreibung
Incremental learning is the subfield of machine learning dealing with continuous adaptation at runtime. In contrast to batch learning the incremental learning problem has a constant low computation- and memory-demand. Hence it is particularly applicable in on-line applications of technical systems where timeliness is an important feature. Furthermore for very large data-sets incremental learning is important especially because of the lowered memory demand.
The seminar deals with different solutions to the incremental learning problem, e.g. Tikhonov regularization or gradient descent methods, and compares the approaches as well for regression estimation as for pattern recognition.
Every student prepares a talk on a subject in the field of incremental learning and in the subsequent discussion the relationship to other talks will be developed, hence covering the problems and ideas of incremental learning altogether.
Studienbereiche
- Informatik > Master of Science in Informatik (bis PO 2016)
- Informatik > Seminare
Modulzuordnungen der Veranstaltung
- Master of Science Wirtschaftsinformatik > INF-MAS2 - Masterseminar 2
