Adaptive Volumetric Mechanical Property Fields Invariant to Resolution

Rishit Dagli, Donglai Xiang, Vismay Modi, Xuning Yang, Gavriel State, David Levin I.W., Maria Shugrina
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:22286-22333, 2026.

Abstract

Accurate mechanical properties (or materials) Young’s modulus ($E$), Poisson’s ratio ($\nu$) and density ($\rho$) are essential for reliable physics simulation of digital worlds, but most 3D assets lack this information. We propose AdaVoMP, a method for predicting accurate dense spatially-varying $(E, \nu, \rho)$ for input 3D objects across representations, improving the resolution, accuracy, and memory efficiency over the state-of-the-art. The foundation of our technique is a sparse and adaptive voxel structure SAV that efficiently represents both the input 3D shape and the material field output. We replace the fixed-voxel model of the most accurate prior method, VoMP, with a novel sparse transformer encoder-decoder model that learns to generate a unique SAV autoregressively for every input shape to represent its materials, achieving a resolution $16^3\times$ higher than prior art. Experiments show that AdaVoMP estimates more accurate volumetric properties, even with lesser test-time compute than all prior art. This allows us to convert high-resolution complex 3D objects into simulation-ready assets, resulting in realistic deformable simulations.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-dagli26a, title = {Adaptive Volumetric Mechanical Property Fields Invariant to Resolution}, author = {Dagli, Rishit and Xiang, Donglai and Modi, Vismay and Yang, Xuning and State, Gavriel and I.W., David Levin and Shugrina, Maria}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {22286--22333}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/dagli26a/dagli26a.pdf}, url = {https://proceedings.mlr.press/v306/dagli26a.html}, abstract = {Accurate mechanical properties (or materials) Young’s modulus ($E$), Poisson’s ratio ($\nu$) and density ($\rho$) are essential for reliable physics simulation of digital worlds, but most 3D assets lack this information. We propose AdaVoMP, a method for predicting accurate dense spatially-varying $(E, \nu, \rho)$ for input 3D objects across representations, improving the resolution, accuracy, and memory efficiency over the state-of-the-art. The foundation of our technique is a sparse and adaptive voxel structure SAV that efficiently represents both the input 3D shape and the material field output. We replace the fixed-voxel model of the most accurate prior method, VoMP, with a novel sparse transformer encoder-decoder model that learns to generate a unique SAV autoregressively for every input shape to represent its materials, achieving a resolution $16^3\times$ higher than prior art. Experiments show that AdaVoMP estimates more accurate volumetric properties, even with lesser test-time compute than all prior art. This allows us to convert high-resolution complex 3D objects into simulation-ready assets, resulting in realistic deformable simulations.} }
Endnote
%0 Conference Paper %T Adaptive Volumetric Mechanical Property Fields Invariant to Resolution %A Rishit Dagli %A Donglai Xiang %A Vismay Modi %A Xuning Yang %A Gavriel State %A David Levin I.W. %A Maria Shugrina %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-dagli26a %I PMLR %P 22286--22333 %U https://proceedings.mlr.press/v306/dagli26a.html %V 306 %X Accurate mechanical properties (or materials) Young’s modulus ($E$), Poisson’s ratio ($\nu$) and density ($\rho$) are essential for reliable physics simulation of digital worlds, but most 3D assets lack this information. We propose AdaVoMP, a method for predicting accurate dense spatially-varying $(E, \nu, \rho)$ for input 3D objects across representations, improving the resolution, accuracy, and memory efficiency over the state-of-the-art. The foundation of our technique is a sparse and adaptive voxel structure SAV that efficiently represents both the input 3D shape and the material field output. We replace the fixed-voxel model of the most accurate prior method, VoMP, with a novel sparse transformer encoder-decoder model that learns to generate a unique SAV autoregressively for every input shape to represent its materials, achieving a resolution $16^3\times$ higher than prior art. Experiments show that AdaVoMP estimates more accurate volumetric properties, even with lesser test-time compute than all prior art. This allows us to convert high-resolution complex 3D objects into simulation-ready assets, resulting in realistic deformable simulations.
APA
Dagli, R., Xiang, D., Modi, V., Yang, X., State, G., I.W., D.L. & Shugrina, M.. (2026). Adaptive Volumetric Mechanical Property Fields Invariant to Resolution. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:22286-22333 Available from https://proceedings.mlr.press/v306/dagli26a.html.

Related Material