Davis Rempe

Stanford University, Nvidia (United Kingdom)

Papers

Total Citations

H-Index

About

Davis Rempe is a leading researcher in computer vision and embodied AI, whose work bridges the gap between physical world understanding and machine intelligence. His primary research areas include 3D scene understanding, physical dynamics prediction, and human-environment interaction modeling. Rempe’s major contributions lie in enabling machines to anticipate the physical behavior of objects and humans in real-world settings. His seminal work on predicting the dynamics of unseen 3D objects (7 citations) laid the foundation for creating robots and virtual worlds that can reason about physical interactions, while his earlier research on generalizable rigid object dynamics (3 citations) established core principles for learning physical laws from visual data. Notably, his recent COPILOT framework (5 citations) introduces the novel problem of collision prediction from egocentric video, a breakthrough for VR, AR, and assistive robotics safety. By combining geometric reasoning with data-driven learning, Rempe’s work has significant implications for autonomous systems that must navigate and interact with dynamic environments. His research continues to push the boundaries of how machines perceive and predict physical reality.