AtomWorld-Mem: Memory-Restored World States for Long-Horizon Atomistic Evolution
Abstract
High-fidelity atomistic evolution over long timescales requires more than observing the current crystal configuration. Instantaneous atomistic snapshots are often incomplete: locally similar configurations can correspond to different hidden dynamical contexts, future event preferences, and waiting-time scales. We argue that this snapshot ambiguity makes long-horizon atomistic evolution fundamentally a memory-based world-state restoration problem. To address this, we introduce AtomWorld-Mem, a memory-restored atomistic world model that recovers the latent world state missing from instantaneous crystal snapshots. AtomWorld-Mem treats the evolving alloy as an AtomWorld: spatial encoders write multi-scale atomistic keyframes from dense local topology and sparse long-range defect context, while short-term event memory and long-term structural memory integrate these keyframes across time to restore a future-predictive evolutionary state. The restored state is used to prioritize legal vacancy-mediated events under single-event Kinetic Monte Carlo (KMC) constraints, while event legality, physical execution, and residence-time updates remain governed by the underlying simulator. Empirically, AtomWorld-Mem improves long-horizon atomistic progress under fixed microscopic event budgets while maintaining high-fidelity evolution across energetic, structural, and vacancy-transport observables. It further transfers zero-shot across diverse unseen alloy-temperature AtomWorlds, suggesting that the learned memory-restoration mechanism captures reusable principles of hidden-state inference rather than a system-specific local energy heuristic. These results position memory-restored world-state modeling as a promising route toward efficient, physically grounded, and transferable atomistic evolution.
Related
- AtomWorld-Mirror: Macro-Step World Modeling of Critical Evolution Backbones for Materials Dynamics
- Atomic memory in molecular charge evolution in deep inner-shell Auger cascades.
- Fragmented and symmetry-restored condensates in a three-site Bose ring: exchange versus umklapp
- Atomistic dynamics of early oxidation of Si upon ultra-short pulsed laser heating
- Aitomia: An Agentic Framework for AI-Driven Atomistic and Quantum Chemical Simulations
BibTeX
@misc{luo2026atomworldmem,
title = {{AtomWorld-Mem: Memory-Restored World States for Long-Horizon Atomistic Evolution}},
author = {Tian Luo and Ruge Zhang and Haozhi Han and Yifeng Chen and Yunquan Zhang and Yunxin Liu and Ting Cao and Kun Li},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2609.31133},
eprint = {2609.31133},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.31133},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - AtomWorld-Mem: Memory-Restored World States for Long-Horizon Atomistic Evolution
AU - Tian Luo
AU - Ruge Zhang
AU - Haozhi Han
AU - Yifeng Chen
AU - Yunquan Zhang
AU - Yunxin Liu
AU - Ting Cao
AU - Kun Li
PY - 2026
DA - 2026/10/20
PB - arXiv
DO - 10.48550/arxiv.2609.31133
UR - https://arxiv.org/abs/2609.31133
AB - High-fidelity atomistic evolution over long timescales requires more than observing the current crystal configuration. Instantaneous atomistic snapshots are often incomplete: locally similar configurations can correspond to different hidden dynamical contexts, future event preferences, and waiting-time scales. We argue that this snapshot ambiguity makes long-horizon atomistic evolution fundamentally a memory-based world-state restoration problem. To address this, we introduce AtomWorld-Mem, a memory-restored atomistic world model that recovers the latent world state missing from instantaneous crystal snapshots. AtomWorld-Mem treats the evolving alloy as an AtomWorld: spatial encoders write multi-scale atomistic keyframes from dense local topology and sparse long-range defect context, while short-term event memory and long-term structural memory integrate these keyframes across time to restore a future-predictive evolutionary state. The restored state is used to prioritize legal vacancy-mediated events under single-event Kinetic Monte Carlo (KMC) constraints, while event legality, physical execution, and residence-time updates remain governed by the underlying simulator. Empirically, AtomWorld-Mem improves long-horizon atomistic progress under fixed microscopic event budgets while maintaining high-fidelity evolution across energetic, structural, and vacancy-transport observables. It further transfers zero-shot across diverse unseen alloy-temperature AtomWorlds, suggesting that the learned memory-restoration mechanism captures reusable principles of hidden-state inference rather than a system-specific local energy heuristic. These results position memory-restored world-state modeling as a promising route toward efficient, physically grounded, and transferable atomistic evolution.
ER -