Incremental Potential Contact (IPC) is for robust and accurate time stepping of nonlinear elastodynamics. IPC guarantees intersection- and inversion-free trajectories regardless of materials, time-step sizes, velocities, or deformation severity.

Code for the paper "Thermodynamics-informed graph neural networks" published in IEEE Transactions on Artificial Intelligence (TAI).

IsoGCN code for ICLR2021

PENN code for NeurIPS 2022

Open source implemention of Moving Particle Semi-implcit (MPS) method

Reconstruction and Simulation of Elastic Objects with Spring-Mass 3D Gaussians

[NeurIPS 2022] The implementation for the paper "Learning Physical Dynamics with Subequivariant Graph Neural Networks".

An extensible benchmark suite to evaluate data-driven physical simulation

Code for the paper "Structure-preserving neural networks" published in Journal of Computational Physics (JCP).

[ICLR24] CinDM uses compositional generative models to design boundaries and initial states significantly more complex than the ones seen in training for physical simulation

Code for the paper "Deep learning of thermodynamics-aware reduced-order models from data" published in Computer Methods in Applied Mechanics and Engineering (CMAME).

mpm_solver

Applications of the Teg differentiable programming language to problems spanning graphics and physical simulation.

Source files for my experiments not limited to computer graphics.

Implementation of CogSci 2019 paper 'Active physical learning via reinforcement learning'

[AAAI24] LE-PDE-UQ endows deep learning-based surrogate models with robust and efficient uncertainty quantification capabilities for both forward and inverse problems.

This is our course homework. I create a physical simulation program to demonstrate the dynamics of human hair under the influence of air flow.

Visualize different modules related to motor control and operations.

Simulink model and python interface to simulate electrical motor operations.

Coursework for my Y3 Parallel Scientific Computing I module! Working on parallel physical simulations in C++ using OpenMP for parallelisation.