MOSAIC: Addressing Space for Local, Steerable, and Scalable Atomistic Generation
Abstract
Generative models are opening new possibilities in molecular discovery and materials science. Realizing this potential, however, requires both local control over atomic placement and the ability to assemble large, heterogeneous structures. We introduce Mosaic, a hybrid mesh-particle representation that partitions space into anchored regions containing atom slots. Each slot encodes occupancy, atomic species, and a continuous displacement from its anchor, providing persistent spatial addresses for local conditioning, hierarchical generation, and scalable assembly. Because Mosaic defines a representation rather than a model, it can be paired with different generative backbones. With the molecular backbone held fixed, its local support and coordinate prior improve molecular connectivity and validity, the resulting model samples efficiently across molecular sizes and enables fine-grained spatial control. At larger material scales, the same representation enables generation of atomistic polycrystals conditioned on coarse spatial layouts. Most notably, locality turns large-scale generation into a bounded assembly problem, enabling structure generation with more than one million atoms. Together, this establishes local spatial addressing as a practical foundation for steerable generation of matter across scales, from individual molecules to large-scale structures of broader engineering significance.
Related
- Scalable long-range machine learning interatomic potentials with electrical response for atomistic material modeling
- GLASS: Global Latent Aggregation with Slot-based Set Decoding for Scalable All-Atom Crystal Generation
- Deterministic and Programmable Fusion for Scalable Generation of Photonic Graph States
- Steerable Radiation Forces with Frequency-Detuned Acoustic Metasurfaces
- Topologically steerable ultra-flat electro-optic combs via non-Hermitian frequency lattices
BibTeX
@misc{nadkarni2026mosaic,
title = {{MOSAIC: Addressing Space for Local, Steerable, and Scalable Atomistic Generation}},
author = {Ishan Nadkarni and Sebastien Hamel and Flynn Walsh and Fei Zhou},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.04161},
eprint = {2610.04161},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.04161},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - MOSAIC: Addressing Space for Local, Steerable, and Scalable Atomistic Generation
AU - Ishan Nadkarni
AU - Sebastien Hamel
AU - Flynn Walsh
AU - Fei Zhou
PY - 2026
DA - 2026/10/03
PB - arXiv
DO - 10.48550/arxiv.2610.04161
UR - https://arxiv.org/abs/2610.04161
AB - Generative models are opening new possibilities in molecular discovery and materials science. Realizing this potential, however, requires both local control over atomic placement and the ability to assemble large, heterogeneous structures. We introduce Mosaic, a hybrid mesh-particle representation that partitions space into anchored regions containing atom slots. Each slot encodes occupancy, atomic species, and a continuous displacement from its anchor, providing persistent spatial addresses for local conditioning, hierarchical generation, and scalable assembly. Because Mosaic defines a representation rather than a model, it can be paired with different generative backbones. With the molecular backbone held fixed, its local support and coordinate prior improve molecular connectivity and validity, the resulting model samples efficiently across molecular sizes and enables fine-grained spatial control. At larger material scales, the same representation enables generation of atomistic polycrystals conditioned on coarse spatial layouts. Most notably, locality turns large-scale generation into a bounded assembly problem, enabling structure generation with more than one million atoms. Together, this establishes local spatial addressing as a practical foundation for steerable generation of matter across scales, from individual molecules to large-scale structures of broader engineering significance.
ER -