A Neural-preconditioned Poisson Solver for Mixed Dirichlet and Neumann Boundary Conditions

no code implementations • 29 Sep 2023 • Kai Weixian Lan, Elias Gueidon, Ayano Kaneda, Julian Panetta, Joseph Teran

The core of our solver is a neural network trained to approximate the inverse of a discrete structured-grid Laplace operator for a domain of arbitrary shape and with mixed boundary conditions.

A Deep Conjugate Direction Method for Iteratively Solving Linear Systems

no code implementations • 22 May 2022 • Ayano Kaneda, Osman Akar, Jingyu Chen, Victoria Kala, David Hyde, Joseph Teran

We present a novel deep learning approach to approximate the solution of large, sparse, symmetric, positive-definite linear systems of equations.

Analytically Integratable Zero-restlength Springs for Capturing Dynamic Modes unrepresented by Quasistatic Neural Networks

no code implementations • 25 Jan 2022 • Yongxu Jin, Yushan Han, Zhenglin Geng, Joseph Teran, Ronald Fedkiw

We present a novel paradigm for modeling certain types of dynamic simulation in real-time with the aid of neural networks.

A Momentum-Conserving Implicit Material Point Method for Surface Energies with Spatial Gradients

no code implementations • 29 Jan 2021 • Jingyu Chen, Victoria Kala, Alan Marquez-Razon, Elias Gueidon, David A. B. Hyde, Joseph Teran

We present a novel Material Point Method (MPM) discretization of surface tension forces that arise from spatially varying surface energies.

Graphics Computational Engineering, Finance, and Science I.3.0; I.6.0