DMIS: Dynamic Mesh-based Importance Sampling for Training Physics-Informed Neural Networks

Zijiang Yang, Zhongwei Qiu, Dongmei Fu

Modeling dynamics in the form of partial differential equations (PDEs) is an effectual way to understand real-world physics processes. For complex physics systems, analytical solut…