Data Party THU

Sep 26, 2026 · Artificial Intelligence

MIT's GeoPT Halves Physics Simulation Data Needs, Doubles Convergence via Geometric Pre-Training

MIT CSAIL and Tsinghua researchers introduce GeoPT, a geometric pre-training method that uses synthetic dynamics on 10,000+ 3D shapes to accelerate physics simulation convergence by 2x and reduce required training data by 20-60% across aerodynamics, hydrodynamics, and collision tasks.