How does the human brain learn to see so quickly—and what can that teach us about building smarter AI? This question is at the heart of Mastering Vision: How Inductive Biases Shape Mammalian Learning Efficiency, a new project recently awarded funding in the Open Competition Domain Science-M programme, led by Sander Bohte and Steven Scholte (both CWI & University of Amsterdam).
While artificial intelligence systems rely on enormous datasets to learn visual tasks, the human brain requires far less information to make sense of the world. The key lies in inductive biases—built-in assumptions that guide the brain’s interpretation of visual input. This project explores how these biases develop from infancy to adulthood, enabling rapid and efficient learning in the mammalian visual system.
By combining computational modeling with neuroscience data, the researchers aim to uncover the principles behind this natural learning efficiency. The findings could reshape our understanding of visual cognition, inform the design of more data-efficient AI, and potentially guide clinical approaches for neurodevelopmental and perceptual disorders.
The research is funded by the Netherlands Organisation for Scientific Research (NWO) as part of the Open Competition Domain Science-M programme, which supports curiosity-driven fundamental research across scientific disciplines.
About the researchers
Prof dr. Sander Bohte leads the Machine Learning group at Centrum Wiskunde & Informatica (CWI). His work focuses on biologically inspired neural networks, computational neuroscience, and efficient AI algorithms. Bohte is known for his contributions to spiking neural networks and brain-inspired learning architectures.
Dr. Steven Scholte is an associate professor at the University of Amsterdam, where he bridges cognitive neuroscience and artificial intelligence. His research explores the neural basis of perception and learning, often combining neuroimaging with computational modeling to understand how the brain processes complex information.
Picture: Shutterstock
