The behavior of species (i.,e. classes of similar natural agents) is optimized on different time-scales, ranging from long-term optimization of morphology and structure on the level of generations of agents (evolution) to short-term optimization of an agents behavior during its lifetime (learning). Even though evolution and learning happen on different time-scales, they are not independent. In contrast, learning and evolution can support each other in finding appropriate solutions for the overall problem. The talk will discuss how evolution and learning can be combined to create efficient, large-scale artificial neural networks, which can be used e.,g. for robotic control tasks or as a component of a learning architecture. One key concept in this context is development, i.,e. the mapping of genotypes onto phenotypes, since it connects the different representations on which learning and evolution act.
On the Interplay of Learning and Evolution
VeranstaltungsortRaum Seminarraum 117, Robert-Hooke-Str. 5 in Bremen
In der Regel sind die Vorträge Teil von Lehrveranstaltungsreihen der Universität Bremen und nicht frei zugänglich. Bei Interesse wird um Rücksprache mit dem Sekretariat unter sek-ric(at)dfki.de gebeten.