PhysiFormer: Learning to Simulate Mechanics in World Space

Hugging Face Daily Papers Papers
physics-simulation diffusion-transformers 3d-meshes world-coordinates motion-generation multi-object-reasoning vertex-trajectories

Summary

PhysiFormer uses coordinate-space diffusion to generate physically-plausible 3D object motions without explicit inductive biases, enabling efficient multi-object reasoning and generalization to complex materials and geometries.

We present PhysiFormer, a diffusion transformer for physically-plausible 3D object motion. Unlike video world models that operate in view-dependent pixel space, PhysiFormer represents objects as 3D meshes expressed in world coordinates. Given the initial vertex positions and velocities, as well as object material type, rigid or elastic, the model samples future vertex trajectories. While related neural physics approaches build on ad-hoc latent spaces or explicitly enforce rigidity and causality, PhysiFormer shows that excellent results can be obtained without any such inductive biases, by casting vertex trajectory prediction as a single denoising diffusion process directly in world coordinates. The probabilistic formulation captures uncertainty in the learned dynamics, enabling diverse plausible futures from initial conditions, making this framework potentially useful for applications with unobserved uncertainty. The model features attention factorised over time, space, and objects for efficiency, enabling permutation-invariant multi-object reasoning without needing explicit object encoding. Trained on over 100k simulated trajectories, PhysiFormer generates rigid and elastic mechanics, and generalises to mixed-material settings, unseen real-world geometries, and larger object counts. It substantially outperforms autoregressive baselines in trajectory accuracy, rigidity preservation, and momentum-based physical consistency. Our results position coordinate-space diffusion as a promising step toward view-invariant, geometry-aware world modelling for robotics, graphics, and physical design. Visualisations, code, and models are available at https://yimingc9.github.io/physiformer.
Original Article
View Cached Full Text

Cached at: 06/26/26, 02:04 AM

Paper page - PhysiFormer: Learning to Simulate Mechanics in World Space

Source: https://huggingface.co/papers/2606.27364

Abstract

PhysiFormer uses coordinate-space diffusion to generate physically-plausible 3D object motions without explicit inductive biases, enabling efficient multi-object reasoning and generalization to complex materials and geometries.

We present PhysiFormer, adiffusion transformerfor physically-plausible 3D object motion. Unlike video world models that operate in view-dependent pixel space, PhysiFormer represents objects as3D meshesexpressed inworld coordinates. Given the initial vertex positions and velocities, as well as object material type, rigid or elastic, the model samples futurevertex trajectories. While related neural physics approaches build on ad-hoc latent spaces or explicitly enforce rigidity and causality, PhysiFormer shows that excellent results can be obtained without any such inductive biases, by casting vertex trajectory prediction as a singledenoising diffusion processdirectly inworld coordinates. Theprobabilistic formulationcaptures uncertainty in the learned dynamics, enabling diverse plausible futures from initial conditions, making this framework potentially useful for applications with unobserved uncertainty. The model featuresattention factorisedover time, space, and objects for efficiency, enablingpermutation-invariantmulti-object reasoning without needing explicit object encoding. Trained on over 100k simulated trajectories, PhysiFormer generates rigid and elastic mechanics, and generalises to mixed-material settings, unseen real-world geometries, and larger object counts. It substantially outperformsautoregressive baselinesin trajectory accuracy, rigidity preservation, and momentum-basedphysical consistency. Our results position coordinate-space diffusion as a promising step toward view-invariant, geometry-aware world modelling for robotics, graphics, and physical design. Visualisations, code, and models are available at https://yimingc9.github.io/physiformer.

View arXiv pageView PDFAdd to collection

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2606.27364 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2606.27364 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2606.27364 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

PhysisForcing: Physics Reinforced World Simulator for Robotic Manipulation

Hugging Face Daily Papers

PhysisForcing is a training framework that enhances embodied video generation for robotic manipulation by enforcing physical consistency through pixel-level trajectory alignment and semantic-level relational alignment losses in a DiT-based architecture, achieving notable improvements on benchmarks.

Stable and Counterfactually Robust Physical World Models from Imposed Structure and Learned Physics

arXiv cs.LG

This paper proposes a physics-informed world model that generates dynamics from the gradient of a learned energy via a fixed reversible operator, with one-way energy dissipation and structured parameter interventions. Across electromagnetic, particle-in-cell, and shallow-water systems, ~9k-parameter models achieve stable 100x-horizon rollouts, correct counterfactual responses, and out-of-distribution refusal when no intervention map is supplied, far outperforming unstructured equal-capacity baselines.