Physics-Editable World Models: Data, Interventions, and Evaluation for Controllable Simulation
PDF

Keywords

Physics-Editable World Models
Dataset Design
Physical Factors
Counterfactual Editing
Temporal Consistency
Evaluation

Abstract

Physics-Editable World Models now turns on the ability to balance performance with evidence quality, resource limits, and transfer across settings. This technical review examines distinguishing visually plausible editing from interventions that preserve physical consistency. The comparison integrates 1 focal paper with 11 independently retrieved publications verified through persistent DOI or publisher records. The analysis is organized around dataset design, physical factors, counterfactual editing, temporal consistency, and evaluation. A central precaution is not to treat results from heterogeneous studies as exchangeable, the review compares problem formulation, design assumptions, and transfer boundary. Across the literature, the comparison suggests that advances in physics-editable world models become credible when representation, objective, and evaluation protocol are evaluated together and when uncertainty about distribution shift is reported explicitly. The synthesis ties method selection to use-case risk and reveals common limits on generalization, and proposes a research agenda centered on transparent baselines, stress testing, and reproducible evidence.

PDF

References

Hu, B., Ma, Y., Huang, J., Zhang, Z., Wu, H., Zhang, R., ... & Li, X. (2026). PhysEditWorld: A Large-Scale Dataset Toward Physics-Editable World Models. arXiv preprint arXiv:2606.26694.

He, Y., Yuan, Z., Tu, Z., Ye, Y., & Sun, L. (2026). 3D4D: An Interactive, Editable, 4D World Model via 3D Video Generation. Proceedings of the AAAI Conference on Artificial Intelligence, 40(48), 41595-41597. https://doi.org/10.1609/aaai.v40i48.42351

Chen, J., Xiao, G., Han, X., & Chen, H. (2021). Controllable and Editable Neural Story Plot Generation via Control-and-Edit Transformer. IEEE Access, 9, 96692-96699. https://doi.org/10.1109/access.2021.3094263

Zhao, Y., Li, H., Zhang, Z., Chen, Y., Liu, Q., & Zhang, X. (2024). Regional Traditional Painting Generation Based on Controllable Disentanglement Model. IEEE Transactions on Circuits and Systems for Video Technology, 34(8), 6913-6925. https://doi.org/10.1109/tcsvt.2023.3298811

Ananthakrishnan,, Bharathi, A., Dharanivendhan,, & Muthuganapathy, R. (2025). CAD-RAG: A multi-modal retrieval augmented framework for user editable 3D CAD model generation. Journal of WSCG, 33(1-2). https://doi.org/10.24132/jwscg.2025-11

Yang, C., Wang, C., Liang, L., & Su, Z. (2024). Real-world image denoising via efficient diffusion model with controllable noise generation. Journal of Electronic Imaging, 33(04). https://doi.org/10.1117/1.jei.33.4.043003

Arnaud, A., Gouiffès, M. l., & Ammi, M. (2022). Towards Real-Time 3D Editable Model Generation for Existing Indoor Building Environments on a Tablet. Frontiers in Virtual Reality, 3. https://doi.org/10.3389/frvir.2022.782564

Thompson, D. (2020). Real-world evidence generation with David Thompson. Video Journal of Biomedicine. https://doi.org/10.2217/vjbm-2020-0001

Zhang, H., Shi, P., & Yang, G. (2026). TDtoon: Two-stage diffusion for controllable cartoon video generation via sketch enhancement. Displays, 91, 103269. https://doi.org/10.1016/j.displa.2025.103269

Zhu, Y., Li, W., Fan, C., & Yu, L. (2026). AIGC-Driven Short Video Generation Based on the Controllable Multimodal Fusion Architecture. Electronics, 15(9), 1783. https://doi.org/10.3390/electronics15091783

Hu, Y., Luo, C., & Chen, Z. (2024). A Benchmark for Controllable Text -Image-to-Video Generation. IEEE Transactions on Multimedia, 26, 1706-1719. https://doi.org/10.1109/tmm.2023.3284989

Sah, M., Mathew, J., & Dayananda, P. (2026). A real-world framework for automated product recognition and catalog generation: dataset, model, and analysis. Scientific Reports, 16(1). https://doi.org/10.1038/s41598-026-42266-9