President of the Czech Republic Petr Pavel appointed new university professors on June 1, 2026, during a ceremony held in the Great Hall of the Carolinum. Among the ninety-three newly appointed professors was Josef Šivic from the Czech Institute of Informatics, Robotics and Cybernetics at CTU (CIIRC CTU), who was appointed Professor in the field of Informatics and Artificial Intelligence upon the recommendation of the Scientific Council of the Czech Technical University in Prague.
Professor Josef Šivic is one of the leading Czech experts in computer vision, machine learning, and artificial intelligence. His research bridges robotics and biology by developing AI models capable of understanding space and physical laws, whether in enabling robots to reliably perceive their physical surroundings or in analyzing complex biomolecular structures.
At CIIRC, Josef Šivic serves as a senior research scientist and leads a research group focused on machine perception and artificial intelligence. He is the recipient of a prestigious ERC Advanced Grant, Director of ELLIS Unit Prague, which he is expanding into the nationwide ELLIS Unit Czechia initiative, and head of the open initiative AICzechia. In 2023, he received the Czech AI Award for research and development, and in 2025 he was awarded the Ordre des Palmes Académiques, a distinction recognizing contributions to education and scientific cooperation with France.
His inaugural professorial lecture before the Scientific Council of the Faculty of Electrical Engineering was titled “Automatic Visual Recognition: From Pixels to Machines That See.” In the lecture, he presented his vision for the future development of computer vision and artificial intelligence. He emphasized that while current systems achieve extraordinary results in processing images and text, the real world is three-dimensional, dynamic, and governed by physical laws. According to him, the future lies in so-called embodied perception, meaning the ability of machines to learn to perceive, understand, and influence the world through observation and interaction.
Among other topics, his lecture presented research focused on using millions of instructional videos to teach robotic manipulation skills with minimal human supervision. Another part addressed the development of visual representations that incorporate geometric and physical principles, enabling robots to better understand three-dimensional motion and acting forces. He also demonstrated how similar methods can be applied to the study of biomolecular structures and the prediction of how mutations affect protein interactions, which could in the future contribute to the development of treatments for neurodegenerative and other diseases.
More about Prof. Josef Šivic:
