Dr. Elahe Arani (World Models Under Drift: Generative Foundations for Reliable Embodied AI)
Elahe Arani
Abstract
Distribution shift is a core challenge for embodied AI deployed in the real world. Generative world models offer a scalable framework for modeling dynamic environments and supporting simulation, reasoning, and policy learning. This perspective suggests a unified approach to robustness and generalization in non-stationary settings.
Speaker
Elahe Arani
Video
Chat is not available.
Successful Page Load