Next generation machine learning
Artificial neural networks are inspired by biological brains, but their conceptual foundations are quite old: they were modeled after the brain’s anatomy in the 1950s and 1960s – according to the understanding at that time. Fundamental limitations remain, in particular, the need for large amounts of training data and the difficulties in generalizing and extrapolating between domains.
A deeper mathematical understanding of cognitive processes like auditory or visual perception may trigger progress both in neuroscience and in artificial intelligence. Project proposals will be considered that seek to overcome current limitations in AI by a new generation of algorithms, inspired by today’s neurosciences and by advances in brain research.