Title
Core representational dimensions of visually-perceived objects
Visual perception of our world seems incredibly easy. Yet despite tremendous progress in the field of vision science and recent breakthroughs in artificial intelligence, we still do not know how our visual system allows us to make sense of our world. One reason for this gap in knowledge may lie in the strong focus on trying to understand how our visual system allows us to recognize and categorize objects. Effectively, this approach equates vision with assigning labels to the things surrounding us. In this talk, I will argue that, instead, we should aim much broader at understanding the representational dimensions underlying our ability to make sense of our world. In my talk, I will highlight recent efforts from our group at identifying the core dimensions of visually-perceived objects and how they relate to object coding in the human brain. The results of our work paint a different picture to the traditional view of a hierarchy of the visual system, identifying behaviorally-relevant representations at even the earliest processing stages and indicating that a major organizational principle of the visual system is our ability to reason, communicate, and interact effectively with the visual world.
Biography
Martin Hebart obtained his PhD with John-Dylan Haynes and Tobias Donner at the Bernstein Center for Computational Neuroscience Berlin, followed by a postdoc with Jan Gläscher at the University Medical Center Hamburg-Eppendorf and a Humboldt fellowship with Chris Baker at the National Institute of Mental Health in Bethesda, MD. In 2019, he started the independent research group Vision and Computational Cognition at the Max Planck Institute for Human Cognitive and Brain Sciences in Leipzig. He was recently awarded a European Research Council Starting Grant to investigate the core dimensions underlying visual and semantic representations in the human brain. He is a founder of the THINGS initiative (https://things-initiative.org) for an interdisciplinary, multiscale und multispecies study of object recognition and understanding. His current research is focused on revealing core representational principles in the visual system and how they enable object understanding and effective behavior.