MPMWorlds: MPM Simulations for Inferring and Extrapolating Physical Dynamics
Žiga Kovačič, Kevin Ellis
In submission (2026)
I'm an incoming CS PhD student at Stanford, advised by Jiajun Wu. I'm interested in hierarchical world models that learn the right level of physical abstraction for each task and co-evolve through interaction. I also have a deep interest in physical simulation.
Previously, I completed my undergrad at Cornell, where I was fortunate to work with Kevin Ellis and was recognised as a Merrill Presidential Scholar.
See Google Scholar for the most up-to-date list.
Žiga Kovačič, Kevin Ellis
In submission (2026)
Minchen Li, Chenfanfu Jiang, Zhaofeng Luo, Žiga Kovačič, Wenxin Du, Chang Yu, Tianyi Xie
Online book, v1.0.3 (2026)
Žiga Kovačič*, Justin T. Chiu*, Celine Lee, Wenting Zhao, Kevin Ellis
DL4C @ NeurIPS 2025, extension in submission
Implements Anisotropic Elastoplasticity for Cloth, Knit and Hair Frictional Contact [C. Jiang et al., 2017].
PBD solver that simulates cloth, volumetric meshes, and fluids in the same solver. Code coming soon.
Moving Least Squares Material Point Method, CPIC, various materials, and mesh extraction in 2D and 3D.
Distributive ray tracing, reflections, refractions, volumetrics, emissive objects, and a BVH.
Modifies Gaussian Splatting to remove obstructions blocking objects of interest in 3D scene reconstruction. Code coming soon.
Generates node-edge graph programs from sketches using vision-language models and differential evolution.
Implements "Image-Space Modal Bases for Plausible Manipulation of Objects in Video", extended with modal warping.