Understanding how humans develop rich, structured cognition from infancy remains one of the most profound challenges in science. While neuroscience and psychology have long provided theories of learning and development—emphasizing factors like embodiment, exploration, and grounded language understanding—deep learning models have recently demonstrated surprising success in capturing aspects of human-like behavior, often in the absence of these developmental constraints. This conference will bring together researchers across developmental psychology, cognitive neuroscience, and artificial intelligence to critically examine how insights from human development can inform next-generation machine learning, and vice versa.

We will explore questions such as: (1) What computational principles underlie human developmental trajectories, and can they be formalized in modern AI systems? (2) How do neural and behavioral data from infants and children constrain our models of representation learning? (3) In what ways do current deep learning models fail to capture core aspects of human development—and what might it take to close the gap?