← Back to all news
arXiv Machine Learning By Chukwudalu Dumebi-Kachikwu

GaussianBench: Physics-Fidelity Evaluation for Gaussian Scene Representations

Read the original on arXiv Machine Learning →

GaussianBench is a physics‑fidelity evaluation suite designed for physics‑integrated Gaussian scene representations. It tests key physical properties—such as conservation, continuum response, material coupling, covariance transport, rendering, thermal phase change, and counterfactual response—using frozen scenes, simulator‑independent scorers, and analytical or measured references. The benchmark distinguishes physical failures from unsupported capabilities and invalid comparisons, and the authors also present GaussianFlesh, a thermomechanical reference system that couples rendering primitives with continuum material points via an MPM solver.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

Related stories

arXiv AI
Aug 28

KnockGS:interaction-Grounded Calibrationof Physical Gaussian Representations

The paper introduces KnockGS, a framework that calibrates the elasticity and density of 3D Gaussian objects by analyzing their dynamic response to known forces. By extracting temporal response features from observed dynamics, KnockGS estimates material scales, freezes them into the simulator, and demonstrates that these estimates remain accurate for unseen interactions. Evaluation shows that KnockGS outperforms alternative methods in parameter recovery and response fidelity across diverse material targets.

By Chenchen Ge, Hanwen Shen, Bowen Jing, Jiyuan Cai, Xiaofeng Wang, Hongsen Lei, Weitao Zhou, Dandan Zhang, Haibao Yu
More like this →
arXiv Machine Learning
Oct 28

A Gaussian Process State Space Formulation for Materials Constitutive Modeling: Application to Quasi-brittle Materials

The paper introduces a Gaussian‑process state‑space model (GP‑SSM) for constitutive modeling of quasi‑brittle materials, replacing traditional internal variables with latent states learned from stress‑strain data. The GP‑SSM provides uncertainty quantification, enforces physical constraints, and maps latent states and strain tensors to stress responses via a sparse multi‑output GP. The trained model is implemented in Abaqus/Explicit and validated through four cases, demonstrating accurate reproduction of RVE‑scale histories, confinement effects, unloading behavior, and structural responses comparable to the KCC model.

By Houlin Xu, Chenyang Li, K. T. Ramesh, Michael D. Shields
More like this →