Sami Azirar

Sami Azirar is a Ph.D. student at the Robotics Perception and Learning Lab at the University of Bonn, where he is also associated with the Lamarr Institute for Machine Learning and Artificial Intelligence. His research focuses on training robots using web videos and employing video generation models as world simulators, with interests in world models, trajectory learning, and cross-embodiment transfer. He is the lead author of papers including IQLS, a framework for metadata-driven large language model queries, and SYMBOLIZER, a symbolic model-free task planning approach using vision-language models.