DeformMaster: An Interactive Physics-Neural World Model for Deformable Objects from Videos

Can Li; Jie Gu; Jingmin Chen; Lei Lei; Lei Sun; Ren Li; Zhoujian Li

arxiv: 2605.09586 · v2 · pith:KPNUH5WEnew · submitted 2026-05-10 · 💻 cs.CV · cs.RO

DeformMaster: An Interactive Physics-Neural World Model for Deformable Objects from Videos

Can Li , Zhoujian Li , Ren Li , Jie Gu , Lei Lei , Jingmin Chen , Lei Sun This is my paper

Pith reviewed 2026-05-21 08:38 UTC · model grok-4.3

classification 💻 cs.CV cs.RO

keywords deformable objectsworld modelsphysics simulationneural networksvideo-based modelinginteractive simulationmaterial experts4D appearance

0 comments

The pith

DeformMaster builds an interactive physics-neural model from videos to simulate and render deformable object dynamics under new interactions.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper is trying to establish that real videos of interactions with deformable objects can be turned into an online interactive physics-neural world model. This model unifies dynamics and appearance by preserving physical rollouts augmented by neural residuals, grounding sparse hand motions as distributed actuators, and employing spatially varying constitutive experts for materials. It would matter if true because it enables rolling out future states, novel actions, material changes, and new views with high fidelity from video data alone.

Core claim

DeformMaster is presented as a video-derived interactive physics-neural world model that converts real interaction videos into an online interactive model of deformable objects in a unified dynamics-and-appearance framework. It maintains a structured physical rollout supplemented by a neural residual for unmodeled effects, treats sparse hand motion as a distributed compliant actuator, models material response via spatially varying constitutive experts, and generates high-fidelity 4D appearance driven by the physical predictions. Experiments show it outperforms baselines on real-world sequences for dynamic rollout and rendering while supporting novel interactions and variations.

What carries the argument

Hybrid physics-neural world model using structured physical rollout with neural residual compensation, hand motion as distributed compliant actuator, and spatially varying constitutive experts.

If this is right

Real-world deformable object sequences can be used to train models that accurately roll out future dynamics and render dynamic appearance.
The system supports novel action rollout for unseen interactions.
Material parameters can be varied to simulate different properties.
Dynamic novel-view synthesis is possible from the predicted evolution.
It achieves better performance than state-of-the-art baselines.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

This hybrid method might generalize to approximate physics in other domains like fluid or rigid body simulations.
It could support creation of interactive virtual replicas of physical deformable items captured casually on video.
Further work might explore scaling to scenes with multiple objects or environmental interactions.

Load-bearing premise

The assumption that a neural residual can reliably compensate for unmodeled effects while a structured physical rollout remains dominant, and that sparse hand motion can be grounded as a distributed compliant actuator without introducing large errors in the predicted dynamics.

What would settle it

Providing a video of a deformable object under a new interaction and checking if the model's predicted shape changes and appearances match the real recorded sequence within acceptable error bounds would test the claim.

Figures

Figures reproduced from arXiv: 2605.09586 by Can Li, Jie Gu, Jingmin Chen, Lei Lei, Lei Sun, Ren Li, Zhoujian Li.

**Figure 2.** Figure 2: Overview of DeformMaster. From interaction videos, DeformMaster unifies interactive [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** Figure 3: Qualitative comparison on two representative deformation sequences. Columns progress [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

**Figure 4.** Figure 4: Generalization to novel action-, material-, and view-conditioned prediction with our [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗

**Figure 5.** Figure 5: Qualitative visualization of MoCE and material fields. For readability, the left two panels [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗

**Figure 6.** Figure 6: Ablation summary. Columns (a)–(d) compare four design choices, with dynamics metrics in [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗

**Figure 7.** Figure 7: Online interactive playground for DeformMaster. The interface supports interaction-point [PITH_FULL_IMAGE:figures/full_fig_p013_7.png] view at source ↗

**Figure 8.** Figure 8: Applications of DeformMaster. (a) DeformMaster leverages interactive manipulation to [PITH_FULL_IMAGE:figures/full_fig_p014_8.png] view at source ↗

read the original abstract

World models for deformable objects should recover not only geometry and appearance, but also underlying physical dynamics, interaction grounding, and material behavior. Learning such a model from real videos is challenging because deformable linear, planar, and volumetric objects evolve under high-dimensional deformation, noisy interactions, and complex material response. The model must therefore infer a physical state from visual observations, roll it forward under new interactions, and render the resulting dynamics with high visual fidelity. We present DeformMaster, a video-derived interactive physics-neural world model that turns real interaction videos into an online interactive model of deformable objects within a unified dynamics-and-appearance framework. DeformMaster preserves structured physical rollout while using a neural residual to compensate for unmodeled effects, grounds sparse hand motion as distributed compliant actuator for hand-continuum interaction, represents material response with spatially varying constitutive experts, and drives high-fidelity 4D appearance from the predicted physical evolution. Experiments on real-world deformable-object sequences demonstrate DeformMaster's ability to roll out future dynamics and render dynamic appearance, outperforming state-of-the-art baselines while supporting novel action rollout, material-parameter variation, and dynamic novel-view synthesis. Project page: https://can-lee.github.io/deformmaster-web/

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

DeformMaster combines physics rollout with a neural residual and specific interaction grounding for video-based deformable modeling, but needs clearer checks that the physics stays dominant.

read the letter

The punchline for this paper is that it offers a practical way to create interactive physics-neural models of deformable objects straight from videos, with specific choices for handling interactions and materials. What is actually new is the integration of a structured physical rollout with a neural residual for unmodeled effects, combined with grounding hand motion as a distributed compliant actuator and using spatially varying constitutive experts for material behavior. This setup aims to support novel action rollout and material variation in a unified framework. The paper does well by demonstrating results on real deformable object sequences, where it outperforms state-of-the-art baselines in rolling out dynamics and rendering appearance. The ability to do dynamic novel-view synthesis from the predicted physical evolution is a solid practical feature. Where the soft spots are is in the balance between the physics and neural components. The central claim is that the physical model stays dominant, but as the stress-test notes, nothing explicitly prevents the residual from learning the bulk of the dynamics in complex deformation regimes. If the full paper includes ablations or metrics showing the residual's small contribution, that would address this. Otherwise, the risk is that it functions more like learned video prediction than true physics-based interaction. The circularity burden is moderate because validation needs to be clearly independent of the training videos. This work is for researchers in computer vision and robotics interested in building world models for manipulation and scene generation. A reader focused on hybrid neural-physics methods for deformable objects would find value in the technical choices and the experimental setup. It deserves a serious referee because the problem is relevant and the proposed solution has testable elements like material parameter variation. I recommend engaging with it in peer review, particularly to clarify the evidence for the physics component's role.

Referee Report

2 major / 2 minor

Summary. The paper presents DeformMaster, a video-derived interactive physics-neural world model for deformable objects. It recovers geometry, appearance, physical dynamics, interaction grounding, and material behavior from real interaction videos by preserving a structured physical rollout, adding a neural residual for unmodeled effects, grounding sparse hand motion as a distributed compliant actuator, using spatially varying constitutive experts for material response, and rendering high-fidelity 4D appearance from the predicted state. Experiments on real-world sequences claim outperformance over baselines in future dynamics rollout, novel action generalization, material-parameter variation, and dynamic novel-view synthesis.

Significance. If the central claims hold with the physical component demonstrably dominant, the work would advance interactive world models for deformable objects by providing a unified dynamics-appearance framework grounded in physics yet trainable from video. The support for novel actions and material variation would be a notable strength over pure neural video predictors, with potential impact in robotics simulation and AR/VR.

major comments (2)

[Abstract and §3] Abstract and §3 (Method): The central claim requires that the structured physical rollout (with hand grounding as distributed compliant actuator and constitutive experts) produces the dominant deformation trajectory while the neural residual corrects only small effects. No quantitative evidence is provided, such as residual magnitude norms across sequences or a physics-only ablation, to confirm the physical component remains primary; without this, novel-action rollout risks reducing to in-sample video prediction.
[§4] §4 (Experiments): The reported outperformance lacks supporting details including error bars, dataset statistics, cross-validation splits, or specific metrics for material-parameter variation and novel action rollout. This makes it impossible to assess whether gains are robust or affected by post-hoc choices.

minor comments (2)

[§3.1] Notation for the distributed compliant actuator and constitutive experts should be defined more explicitly with equations to avoid ambiguity in how they integrate with the physical state.
[Figure 4] Figure captions and axis labels in result visualizations could be clarified to directly indicate which sequences correspond to novel actions versus training data.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback. We address each major comment below and have revised the manuscript to incorporate additional quantitative evidence and experimental details.

read point-by-point responses

Referee: [Abstract and §3] Abstract and §3 (Method): The central claim requires that the structured physical rollout (with hand grounding as distributed compliant actuator and constitutive experts) produces the dominant deformation trajectory while the neural residual corrects only small effects. No quantitative evidence is provided, such as residual magnitude norms across sequences or a physics-only ablation, to confirm the physical component remains primary; without this, novel-action rollout risks reducing to in-sample video prediction.

Authors: We agree that explicit quantitative support for the dominance of the physical component strengthens the central claim. In the revised manuscript we have added a physics-only ablation (removing the neural residual) and report per-sequence L2 norms of the residual corrections, showing that residuals remain small relative to the total deformation magnitude. These results confirm that the structured physical rollout, including the distributed compliant actuator and constitutive experts, drives the primary trajectory. revision: yes
Referee: [§4] §4 (Experiments): The reported outperformance lacks supporting details including error bars, dataset statistics, cross-validation splits, or specific metrics for material-parameter variation and novel action rollout. This makes it impossible to assess whether gains are robust or affected by post-hoc choices.

Authors: We acknowledge that additional statistical details improve interpretability. The revised experiments section now includes error bars on all metrics, reports dataset statistics and the cross-validation procedure, and provides quantitative results together with visualizations specifically for material-parameter variation and novel-action rollout to demonstrate robustness. revision: yes

Circularity Check

0 steps flagged

No circularity identified; derivation remains self-contained against external benchmarks

full rationale

The abstract and method framing describe a hybrid physics-neural architecture that preserves a structured physical rollout with a corrective neural residual, grounds hand motion as a distributed actuator, and uses spatially varying constitutive experts. No equations, self-citations, or derivation steps are supplied that reduce a claimed prediction or uniqueness result to a fitted input or prior self-work by construction. The central claim of physics-dominant rollout with residual compensation is presented as an architectural choice rather than a mathematical identity, and the paper positions its contributions as empirically validated on real sequences against baselines. This is the normal case of an independent hybrid model whose validity rests on external data and ablations rather than definitional equivalence.

Axiom & Free-Parameter Ledger

2 free parameters · 1 axioms · 1 invented entities

The central claim rests on several learned neural modules and domain assumptions about physical state inference that are fitted to video data; independent evidence for the invented interaction model is not provided in the abstract.

free parameters (2)

neural residual weights
Learned parameters that compensate for unmodeled physical effects in the rollout.
constitutive expert parameters
Spatially varying material response parameters fitted per region from video observations.

axioms (1)

domain assumption A physical state can be reliably inferred from visual observations alone and rolled forward under new interactions.
Invoked as the foundation for the world-model rollout described in the abstract.

invented entities (1)

distributed compliant actuator no independent evidence
purpose: To model sparse hand motion as a spread-out force for continuum interaction.
Introduced to ground hand-object contact; no independent evidence or falsifiable prediction is stated in the abstract.

pith-pipeline@v0.9.0 · 5764 in / 1492 out tokens · 48744 ms · 2026-05-21T08:38:57.080704+00:00 · methodology

Review history (2 revisions) →

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

We decompose deformation dynamics F_θ,ϕ into a physics block P_θ and a residual block R_ϕ: F_θ,ϕ = P_θ ⊕ R_ϕ. ... ˜s_{t+1} = P_MPM_θ (s_t, a_t) ... Δv_p = R_ϕ(˜s_{t+1}, s_t, h_t)
IndisputableMonolith/Foundation/AlphaCoordinateFixation.lean alpha_pin_under_high_calibration unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

P_mix(F_p; E_p, ν_p) = Σ_k w_{k,p} P_k(F_p; E_p, ν_p) ... experts {NH, Cor, StVK}

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

41 extracted references · 41 canonical work pages

[1]

Particle-Grid Neural Dynamics for Learning Deformable Object Models from

Zhang, Kaifeng and Li, Baoyu and Hauser, Kris and Li, Yunzhu , booktitle =. Particle-Grid Neural Dynamics for Learning Deformable Object Models from. 2025 , url =

work page 2025
[2]

and Su, Hao and Mo, Kaichun and Guibas, Leonidas J

Qi, Charles R. and Su, Hao and Mo, Kaichun and Guibas, Leonidas J. , booktitle =

work page
[3]

and Tancik, Matthew and Barron, Jonathan T

Mildenhall, Ben and Srinivasan, Pratul P. and Tancik, Matthew and Barron, Jonathan T. and Ramamoorthi, Ravi and Ng, Ren , booktitle =

work page
[4]

2025 , url =

Jiang, Hanxiao and Hsu, Hao-Yu and Zhang, Kaifeng and Yu, Hsin-Ni and Wang, Shenlong and Li, Yunzhu , booktitle =. 2025 , url =

work page 2025
[5]

Reconstruction and Simulation of Elastic Objects with Spring-Mass

Zhong, Licheng and Yu, Hong-Xing and Wu, Jiajun and Li, Yunzhu , booktitle =. Reconstruction and Simulation of Elastic Objects with Spring-Mass

work page
[6]

Zhang, Mingtong and Zhang, Kaifeng and Li, Yunzhu , booktitle =. Dynamic. 2024 , url =

work page 2024
[7]

Proceedings of the 37th International Conference on Machine Learning (ICML) , year =

Learning to Simulate Complex Physics with Graph Networks , author =. Proceedings of the 37th International Conference on Machine Learning (ICML) , year =

work page
[8]

2023 , url =

Li, Xuan and Qiao, Yi-Ling and Chen, Peter Yichen and Jatavallabhula, Krishna Murthy and Lin, Ming and Jiang, Chenfanfu and Gan, Chuang , booktitle =. 2023 , url =

work page 2023
[9]

2024 , url =

Xie, Tianyi and Zong, Zeshun and Qiu, Yuxing and Li, Xuan and Feng, Yutao and Yang, Yin and Jiang, Chenfanfu , booktitle =. 2024 , url =

work page 2024
[10]

and Zheng, Changxi and Snavely, Noah and Wu, Jiajun and Freeman, William T

Zhang, Tianyuan and Yu, Hong-Xing and Wu, Rundi and Feng, Brandon Y. and Zheng, Changxi and Snavely, Noah and Wu, Jiajun and Freeman, William T. , booktitle =. 2024 , url =

work page 2024
[11]

2024 , url =

Cai, Junhao and Yang, Yuji and Yuan, Weihao and He, Yisheng and Dong, Zilong and Bo, Liefeng and Cheng, Hui and Chen, Qifeng , booktitle =. 2024 , url =

work page 2024
[12]

2025 , url =

Liu, Zhuoman and Ye, Weicai and Luximon, Yan and Wan, Pengfei and Zhang, Di , booktitle =. 2025 , url =

work page 2025
[13]

Efficient Physics Simulation for

Zhao, Haoyu and Wang, Hao and Zhao, Xingyue and Fei, Hao and Wang, Hongqiu and Long, Chengjiang and Zou, Hua , booktitle =. Efficient Physics Simulation for. 2025 , url =

work page 2025
[14]

2025 , url =

Lin, Yuchen and Lin, Chenguo and Xu, Jianjin and Mu, Yadong , booktitle =. 2025 , url =

work page 2025
[15]

2025 , url =

Chen, Boyuan and Jiang, Hanxiao and Liu, Shaowei and Gupta, Saurabh and Li, Yunzhu and Zhao, Hao and Wang, Shenlong , booktitle =. 2025 , url =

work page 2025
[16]

2025 , url =

Li, Zizhang and Yu, Hong-Xing and Liu, Wei and Yang, Yin and Herrmann, Charles and Wetzstein, Gordon and Wu, Jiajun , booktitle =. 2025 , url =

work page 2025
[17]

2026 , url =

Lv, Chunji and Chen, Zequn and Di, Donglin and Zhang, Weinan and Li, Hao and Chen, Wei and Lei, Yinjie and Li, Changsheng , booktitle =. 2026 , url =

work page 2026
[18]

2025 , url =

Yang, Yu and Zhang, Zhilu and Zhang, Xiang and Zeng, Yihan and Li, Hui and Zuo, Wangmeng , journal =. 2025 , url =

work page 2025
[19]

2026 , url =

Chen, Yunuo and Hu, Yafei and Sun, Lingfeng and Kusnur, Tushar and Herlant, Laura and Jiang, Chenfanfu , journal =. 2026 , url =

work page 2026
[20]

Learning Physics-Grounded

Li, Shiqian and Shen, Ruihong and Ni, Junfeng and Pan, Chang and Zhang, Chi and Zhu, Yixin , booktitle =. Learning Physics-Grounded. 2026 , url =

work page 2026
[21]

2026 , url =

Zhan, Jiahao and Li, Zizhang and Yu, Hong-Xing and Wu, Jiajun , journal =. 2026 , url =

work page 2026
[22]

2026 , url =

Lu, Haoran and Wu, Shang and Zhang, Jianshu and Su, Maojiang and Ye, Guo and Xu, Chenwei and Lu, Lie and Maneriker, Pranav and Du, Fan and Li, Manling and Wang, Zhaoran and Liu, Han , journal =. 2026 , url =

work page 2026
[23]

2026 , url =

Liu, Wei and Chen, Ziyu and Li, Zizhang and Wang, Yue and Yu, Hong-Xing and Wu, Jiajun , journal =. 2026 , url =

work page 2026
[24]

and Wu, Jiajun , booktitle =

Yu, Hong-Xing and Duan, Haoyi and Herrmann, Charles and Freeman, William T. and Wu, Jiajun , booktitle =. 2025 , url =

work page 2025
[25]

Science Robotics , volume =

A Review of Learning-Based Dynamics Models for Robotic Manipulation , author =. Science Robotics , volume =. 2025 , doi =

work page 2025
[26]

2024 , url =

Zhang, Kaifeng and Li, Baoyu and Hauser, Kris and Li, Yunzhu , booktitle =. 2024 , url =

work page 2024
[27]

ACM Transactions on Graphics , volume =

A Moving Least Squares Material Point Method with Displacement Discontinuity and Two-Way Rigid Body Coupling , author =. ACM Transactions on Graphics , volume =. 2018 , doi =

work page 2018
[28]

2025 , url =

Yang, Zhuoyi and Teng, Jiayan and Zheng, Wendi and Ding, Ming and Huang, Shiyu and Xu, Jiazheng and Yang, Yuanming and Hong, Wenyi and Zhang, Xiaohan and Feng, Guanyu and Yin, Da and Gu, Xiaotao and Zhang, Yuxuan and Wang, Weihan and Cheng, Yean and Liu, Ting and Xu, Bin and Dong, Yuxiao and Tang, Jie , booktitle =. 2025 , url =

work page 2025
[29]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , year =

Tora: Trajectory-oriented Diffusion Transformer for Video Generation , author =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , year =

work page
[30]

Advances in Neural Information Processing Systems (NeurIPS) , year =

Force Prompting: Video Generation Models Can Learn and Generalize Physics-based Control Signals , author =. Advances in Neural Information Processing Systems (NeurIPS) , year =

work page
[31]

Goal force: Teaching video models to accomplish physics-conditioned goals.arXiv preprint arXiv:2601.05848,

Goal Force: Teaching Video Models To Accomplish Physics-Conditioned Goals , author =. arXiv preprint arXiv:2601.05848 , year =

work page arXiv
[32]

2024 , url =

Karaev, Nikita and Makarov, Iurii and Wang, Jianyuan and Neverova, Natalia and Vedaldi, Andrea and Rupprecht, Christian , journal =. 2024 , url =

work page 2024
[33]

International Conference on Learning Representations (ICLR) , year =

Depth Anything 3: Recovering the Visual Space from Any Views , author =. International Conference on Learning Representations (ICLR) , year =

work page
[34]

2025 , url =

Wang, Ruicheng and Xu, Sicheng and Yang, Cassie and Yuan, Yue and Tong, Xin and Yang, Jiaolong , booktitle =. 2025 , url =

work page 2025
[35]

Structured

Xiang, Jianfeng and Lv, Zelong and Xu, Sicheng and Deng, Yu and Wang, Ruicheng and Zhang, Bowen and Chen, Dong and Tong, Xin and Yang, Jiaolong , booktitle =. Structured. 2025 , url =

work page 2025
[36]

2020 , url =

Sarlin, Paul-Edouard and DeTone, Daniel and Malisiewicz, Tomasz and Rabinovich, Andrew , booktitle =. 2020 , url =

work page 2020
[37]

2024 , url =

Huang, Yi-Hua and Sun, Yang-Tian and Yang, Ziyi and Lyu, Xiaoyang and Cao, Yan-Pei and Qi, Xiaojuan , booktitle =. 2024 , url =

work page 2024
[38]

ACM Transactions on Graphics , volume =

Embedded Deformation for Shape Manipulation , author =. ACM Transactions on Graphics , volume =. 2007 , doi =

work page 2007
[39]

Grounded

Ren, Tianhe and Liu, Shilong and Jiang, Qing and Wei, Yihao and Zeng, Zhaoyang and Yang, Jing and Liu, Wenlong and Wang, Hao and Liang, Feng and Zhang, Hao and Yang, Lei and Zhang, Lei , year =. Grounded

work page
[40]

Real-to-Sim Robot Policy Evaluation with

Zhang, Kaifeng and Sha, Shuo and Jiang, Hanxiao and Loper, Matthew and Song, Hyunjong and Cai, Guangyan and Xu, Zhuo and Hu, Xiaochen and Zheng, Changxi and Li, Yunzhu , journal =. Real-to-Sim Robot Policy Evaluation with. 2025 , url =

work page 2025
[41]

Hansen, Nikolaus , booktitle =. The. 2006 , doi =

work page 2006

[1] [1]

Particle-Grid Neural Dynamics for Learning Deformable Object Models from

Zhang, Kaifeng and Li, Baoyu and Hauser, Kris and Li, Yunzhu , booktitle =. Particle-Grid Neural Dynamics for Learning Deformable Object Models from. 2025 , url =

work page 2025

[2] [2]

and Su, Hao and Mo, Kaichun and Guibas, Leonidas J

Qi, Charles R. and Su, Hao and Mo, Kaichun and Guibas, Leonidas J. , booktitle =

work page

[3] [3]

and Tancik, Matthew and Barron, Jonathan T

Mildenhall, Ben and Srinivasan, Pratul P. and Tancik, Matthew and Barron, Jonathan T. and Ramamoorthi, Ravi and Ng, Ren , booktitle =

work page

[4] [4]

2025 , url =

Jiang, Hanxiao and Hsu, Hao-Yu and Zhang, Kaifeng and Yu, Hsin-Ni and Wang, Shenlong and Li, Yunzhu , booktitle =. 2025 , url =

work page 2025

[5] [5]

Reconstruction and Simulation of Elastic Objects with Spring-Mass

Zhong, Licheng and Yu, Hong-Xing and Wu, Jiajun and Li, Yunzhu , booktitle =. Reconstruction and Simulation of Elastic Objects with Spring-Mass

work page

[6] [6]

Zhang, Mingtong and Zhang, Kaifeng and Li, Yunzhu , booktitle =. Dynamic. 2024 , url =

work page 2024

[7] [7]

Proceedings of the 37th International Conference on Machine Learning (ICML) , year =

Learning to Simulate Complex Physics with Graph Networks , author =. Proceedings of the 37th International Conference on Machine Learning (ICML) , year =

work page

[8] [8]

2023 , url =

Li, Xuan and Qiao, Yi-Ling and Chen, Peter Yichen and Jatavallabhula, Krishna Murthy and Lin, Ming and Jiang, Chenfanfu and Gan, Chuang , booktitle =. 2023 , url =

work page 2023

[9] [9]

2024 , url =

Xie, Tianyi and Zong, Zeshun and Qiu, Yuxing and Li, Xuan and Feng, Yutao and Yang, Yin and Jiang, Chenfanfu , booktitle =. 2024 , url =

work page 2024

[10] [10]

and Zheng, Changxi and Snavely, Noah and Wu, Jiajun and Freeman, William T

Zhang, Tianyuan and Yu, Hong-Xing and Wu, Rundi and Feng, Brandon Y. and Zheng, Changxi and Snavely, Noah and Wu, Jiajun and Freeman, William T. , booktitle =. 2024 , url =

work page 2024

[11] [11]

2024 , url =

Cai, Junhao and Yang, Yuji and Yuan, Weihao and He, Yisheng and Dong, Zilong and Bo, Liefeng and Cheng, Hui and Chen, Qifeng , booktitle =. 2024 , url =

work page 2024

[12] [12]

2025 , url =

Liu, Zhuoman and Ye, Weicai and Luximon, Yan and Wan, Pengfei and Zhang, Di , booktitle =. 2025 , url =

work page 2025

[13] [13]

Efficient Physics Simulation for

Zhao, Haoyu and Wang, Hao and Zhao, Xingyue and Fei, Hao and Wang, Hongqiu and Long, Chengjiang and Zou, Hua , booktitle =. Efficient Physics Simulation for. 2025 , url =

work page 2025

[14] [14]

2025 , url =

Lin, Yuchen and Lin, Chenguo and Xu, Jianjin and Mu, Yadong , booktitle =. 2025 , url =

work page 2025

[15] [15]

2025 , url =

Chen, Boyuan and Jiang, Hanxiao and Liu, Shaowei and Gupta, Saurabh and Li, Yunzhu and Zhao, Hao and Wang, Shenlong , booktitle =. 2025 , url =

work page 2025

[16] [16]

2025 , url =

Li, Zizhang and Yu, Hong-Xing and Liu, Wei and Yang, Yin and Herrmann, Charles and Wetzstein, Gordon and Wu, Jiajun , booktitle =. 2025 , url =

work page 2025

[17] [17]

2026 , url =

Lv, Chunji and Chen, Zequn and Di, Donglin and Zhang, Weinan and Li, Hao and Chen, Wei and Lei, Yinjie and Li, Changsheng , booktitle =. 2026 , url =

work page 2026

[18] [18]

2025 , url =

Yang, Yu and Zhang, Zhilu and Zhang, Xiang and Zeng, Yihan and Li, Hui and Zuo, Wangmeng , journal =. 2025 , url =

work page 2025

[19] [19]

2026 , url =

Chen, Yunuo and Hu, Yafei and Sun, Lingfeng and Kusnur, Tushar and Herlant, Laura and Jiang, Chenfanfu , journal =. 2026 , url =

work page 2026

[20] [20]

Learning Physics-Grounded

Li, Shiqian and Shen, Ruihong and Ni, Junfeng and Pan, Chang and Zhang, Chi and Zhu, Yixin , booktitle =. Learning Physics-Grounded. 2026 , url =

work page 2026

[21] [21]

2026 , url =

Zhan, Jiahao and Li, Zizhang and Yu, Hong-Xing and Wu, Jiajun , journal =. 2026 , url =

work page 2026

[22] [22]

2026 , url =

Lu, Haoran and Wu, Shang and Zhang, Jianshu and Su, Maojiang and Ye, Guo and Xu, Chenwei and Lu, Lie and Maneriker, Pranav and Du, Fan and Li, Manling and Wang, Zhaoran and Liu, Han , journal =. 2026 , url =

work page 2026

[23] [23]

2026 , url =

Liu, Wei and Chen, Ziyu and Li, Zizhang and Wang, Yue and Yu, Hong-Xing and Wu, Jiajun , journal =. 2026 , url =

work page 2026

[24] [24]

and Wu, Jiajun , booktitle =

Yu, Hong-Xing and Duan, Haoyi and Herrmann, Charles and Freeman, William T. and Wu, Jiajun , booktitle =. 2025 , url =

work page 2025

[25] [25]

Science Robotics , volume =

A Review of Learning-Based Dynamics Models for Robotic Manipulation , author =. Science Robotics , volume =. 2025 , doi =

work page 2025

[26] [26]

2024 , url =

Zhang, Kaifeng and Li, Baoyu and Hauser, Kris and Li, Yunzhu , booktitle =. 2024 , url =

work page 2024

[27] [27]

ACM Transactions on Graphics , volume =

A Moving Least Squares Material Point Method with Displacement Discontinuity and Two-Way Rigid Body Coupling , author =. ACM Transactions on Graphics , volume =. 2018 , doi =

work page 2018

[28] [28]

2025 , url =

Yang, Zhuoyi and Teng, Jiayan and Zheng, Wendi and Ding, Ming and Huang, Shiyu and Xu, Jiazheng and Yang, Yuanming and Hong, Wenyi and Zhang, Xiaohan and Feng, Guanyu and Yin, Da and Gu, Xiaotao and Zhang, Yuxuan and Wang, Weihan and Cheng, Yean and Liu, Ting and Xu, Bin and Dong, Yuxiao and Tang, Jie , booktitle =. 2025 , url =

work page 2025

[29] [29]

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , year =

Tora: Trajectory-oriented Diffusion Transformer for Video Generation , author =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , year =

work page

[30] [30]

Advances in Neural Information Processing Systems (NeurIPS) , year =

Force Prompting: Video Generation Models Can Learn and Generalize Physics-based Control Signals , author =. Advances in Neural Information Processing Systems (NeurIPS) , year =

work page

[31] [31]

Goal force: Teaching video models to accomplish physics-conditioned goals.arXiv preprint arXiv:2601.05848,

Goal Force: Teaching Video Models To Accomplish Physics-Conditioned Goals , author =. arXiv preprint arXiv:2601.05848 , year =

work page arXiv

[32] [32]

2024 , url =

Karaev, Nikita and Makarov, Iurii and Wang, Jianyuan and Neverova, Natalia and Vedaldi, Andrea and Rupprecht, Christian , journal =. 2024 , url =

work page 2024

[33] [33]

International Conference on Learning Representations (ICLR) , year =

Depth Anything 3: Recovering the Visual Space from Any Views , author =. International Conference on Learning Representations (ICLR) , year =

work page

[34] [34]

2025 , url =

Wang, Ruicheng and Xu, Sicheng and Yang, Cassie and Yuan, Yue and Tong, Xin and Yang, Jiaolong , booktitle =. 2025 , url =

work page 2025

[35] [35]

Structured

Xiang, Jianfeng and Lv, Zelong and Xu, Sicheng and Deng, Yu and Wang, Ruicheng and Zhang, Bowen and Chen, Dong and Tong, Xin and Yang, Jiaolong , booktitle =. Structured. 2025 , url =

work page 2025

[36] [36]

2020 , url =

Sarlin, Paul-Edouard and DeTone, Daniel and Malisiewicz, Tomasz and Rabinovich, Andrew , booktitle =. 2020 , url =

work page 2020

[37] [37]

2024 , url =

Huang, Yi-Hua and Sun, Yang-Tian and Yang, Ziyi and Lyu, Xiaoyang and Cao, Yan-Pei and Qi, Xiaojuan , booktitle =. 2024 , url =

work page 2024

[38] [38]

ACM Transactions on Graphics , volume =

Embedded Deformation for Shape Manipulation , author =. ACM Transactions on Graphics , volume =. 2007 , doi =

work page 2007

[39] [39]

Grounded

Ren, Tianhe and Liu, Shilong and Jiang, Qing and Wei, Yihao and Zeng, Zhaoyang and Yang, Jing and Liu, Wenlong and Wang, Hao and Liang, Feng and Zhang, Hao and Yang, Lei and Zhang, Lei , year =. Grounded

work page

[40] [40]

Real-to-Sim Robot Policy Evaluation with

Zhang, Kaifeng and Sha, Shuo and Jiang, Hanxiao and Loper, Matthew and Song, Hyunjong and Cai, Guangyan and Xu, Zhuo and Hu, Xiaochen and Zheng, Changxi and Li, Yunzhu , journal =. Real-to-Sim Robot Policy Evaluation with. 2025 , url =

work page 2025

[41] [41]

Hansen, Nikolaus , booktitle =. The. 2006 , doi =

work page 2006