Integrated Calendar

Understanding and modelling how intelligence emerges from biological mechanisms remains a central challenge in neuroscience and artificial intelligence (AI). In this talk, I explore how key aspects of human visual cognition—learning and guided perception—can be modeled within a biologically inspired computational framework. I present a learning model based on interacting bottom-up and top-down pathways, where feedback signals drive learning through a Counter-Hebbian rule and simultaneously guide perception toward task-relevant information.  This enables an efficient instruction-based visual processing in which neural activity depends on both visual input and task demands. These results provide a concrete mechanism linking learning and attention in biological vision—two processes long believed to be related but rarely unified within a single computational model. They also highlight a gap in current AI vision systems and suggest directions for designing more brain-inspired intelligent systems.

Bio:

Roy Abel is a PhD candidate in the Department of Mathematics and Computer Science at the Weizmann Institute of Science, advised by Prof. Shimon Ullman. His research focuses on biologically inspired models of intelligence, particularly mechanisms of learning and perception in human vision and their implications for artificial intelligence. His work aims to bridge neuroscience and AI by developing computational models that both improve our understanding of the human brain and inspire better intelligent systems. In parallel, he works as a researcher at Imubit, where he focuses on foundation models for time-series data and reinforcement learning.