Learning Physical Simulation with Message Passing Transformer

Zeyi Xu, Yifei Li

Machine learning methods for physical simulation have achieved significant success in recent years. We propose a new universal architecture based on Graph Neural Network, the Messa…