Events

This research mission broadly aims to understand how children grasp new concepts from few examples, how children build upon layers of concepts to reach an understanding of the world and have the flexibility to solve an unbounded range of problems. Can we build AI that starts like a baby and learns like a child?

The goal of building machines that can perceive and act in the world as humans and other animals do has been a focus of AI research efforts for over half a century. Over this same period, neuroscience has sought to achieve a mechanistic understanding of the brain processes underlying perception and action. It stands to reason that these parallel efforts could inform one another. However recent advances in deep learning and transformers have, for the most part, not translated into new neuroscientific insights; and other than deriving loose inspiration from neuroscience, AI has mostly pursued its own course which now deviates strongly from the brain.

The presentation will begin with short introductions to the question from the artificial and natural intelligence perspectives. Then it will dig a bit into three research results: a Bayesian approach to 3D perception, a method for efficient planning under the resulting uncertainty over world models, and a study of the role of spatial cognition in human behavior in a VR object-search task. The presentation will conclude with a discussion.