We study learning in brains and machines.
To teach machines to generalize from examples, we develop algorithms to represent complex signals in meaningful ways. We teach our machines to understand sounds and images, generate them, and reason about them especially in complex scenes and with few samples.
We draw inspiration from biological systems. Brains develop and change following experience. People learn even with only few examples, and can reason and generalize in diverse novel scenarios.
Recent papers and news:
Fast 4D Mesh Generation by Spatio-Temporal Attention Chains, Arxiv, Project Page
A Foundation Model for Continuous Glucose Monitoring Data, Arxiv, Nature 2026
Simulating Clinical Interventions with a Generative Multimodal Model of Human Physiology, Arxiv