Josef Sivic

Short bio

Senior researcher (Directeur de recherche) at Inria, Distinguished researcher at Czech Institute of Informatics, Robotics and Cybernetics, Czech Technical University in Prague. Recipient of an ERC Starting Grant (2013), Sullivan Phd Thesis Prize (2007), and the CVPR (2017) and two ICCV (2017) test-of-time awards. Program chair of ICCV (2015) and associate editor of IJCV and TPAMI journals.

Topics of interest

Computer vision, machine learning, robotics

Project in Prairie

Josef Sivic will address weakly supervised learning and learning visual representations for reasoning, planning, and interacting with complex dynamic environments. He will collaborate with industry partners and contribute to teaching graduate level AI courses. His work will also involve the CIIRC institute in Prague, one of the international partners of PRAIRIE, where he has a part-time appointment.


Imagine, for example, an intelligent production line that automatically learns a new workflow by observing a skilled worker; or a rescue robot that autonomously completes tasks in an otherwise inaccessible environment. Another example is a house helper robot assisting disabled or elderly persons in unforeseen situations. All these applications require machines capable of understanding the changing visual world, learning new skills and adapting them to new environments and unforeseen situations. Building machines that have such capabilities is one of the central problems of artificial intelligence. In my research program, I would like to make a step towards solving these challenges.