Prof. Igor Farkaš presents On bio-inspired neural network models for cognitive robotics
On 2022-06-14 11:00:00 at G205, Karlovo náměstí 13, Praha 2
During the last decade, artificial neural networks have become a dominant
modeling approach in artificial intelligence mainly thank to successful deep
learning. This allows to successfully solve concrete tasks using mostly huge
models and long training times. In the talk, I will instead focus on various
smaller scale models that have also been successfully used as a proof of
concept, e.g. in cognitive robotics but also elsewhere. I will explain
different
learning algorithms such as back-propagation, Universal Bidirectional
Associative Mapping (UBAL), the Self-Organizing Map, or AutoEncoder. One aspect
of all models is their degree of biological relevance which will also be
mentioned.
modeling approach in artificial intelligence mainly thank to successful deep
learning. This allows to successfully solve concrete tasks using mostly huge
models and long training times. In the talk, I will instead focus on various
smaller scale models that have also been successfully used as a proof of
concept, e.g. in cognitive robotics but also elsewhere. I will explain
different
learning algorithms such as back-propagation, Universal Bidirectional
Associative Mapping (UBAL), the Self-Organizing Map, or AutoEncoder. One aspect
of all models is their degree of biological relevance which will also be
mentioned.