Andrej Lúčny
He works at the Department of Applied Informatics, Faculty of Mathematics, Physics and Informatics, Comenius University in Bratislava, as a researcher in the field of cognitive robotics. His research focuses on the integration of artificial intelligence methods into real-time systems, and he has also achieved notable results in the field of computer vision. He is a co-founder of the civic association Robotika.SK, which is dedicated to the popularization of robotics, and a regular judge at the ISTROBOT mobile robot competition. In the past, he served as Head of Systems Development at MicroStep-MIS (monitoring information systems) and as a researcher at MicroEpsilon-Inspection (computer vision for inspection systems).
Sign in to Autumn ITAPA 2026
-
How we control robots and how reliable they are (10 min)
Behind the advanced capabilities of robots are artificial intelligence models based on deep neural networks. The current standard architecture is the transformer, particularly multimodal encoder–decoder architectures. Although these models possess remarkable capabilities, they inherently carry risks that limit their appeal to industry. We will discuss the approaches our research is taking to address these shortcomings.