AtomWorld-Mem: Memory-Restored World States for Long-Horizon Atomistic Evolution
Not on the map. Why?
Abstract (arXiv)
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.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 12 of its 37 references that are in the map. First posted on arXiv 25 Sep 2026.
Builds on
12 of 37 references.
References · 12
- 2019Dream to Control: Learning Behaviors by Latent Imaginationcited by 1,653
- 2018Learning Latent Dynamics for Planning from Pixelscited by 1,264
- 2023Mastering Diverse Domains through World Modelscited by 1,253
- 2020Mastering Atari with Discrete World Modelscited by 1,007
- 2018Recurrent World Models Facilitate Policy Evolutioncited by 785
- 2016Deep Variational Bayes Filters: Unsupervised Learning of State Space Models from Raw Datacited by 156
- 2017DeePMD-kit: A deep learning package for many-body potential energy representation and molecular dynamicscited by 79
- 2021Memory-based Deep Reinforcement Learning for POMDPscited by 54
- 2018Learning and Querying Fast Generative Models for Reinforcement Learningcited by 48
- 2025Partially Observable Reinforcement Learning with Memory Tracescited by 9
- 2018Recurrent Predictive State Policy Networkscited by 6
- 2022From latent dynamics to meaningful representationscited by 5
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026AtomWorld-Mirror: Macro-Step World Modeling of Critical Evolution Backbones for Materials Dynamics
- 2025AtomWorld: A Benchmark for Evaluating Spatial Reasoning in Large Language Models on Crystalline Materials
- 2026Atomistic Language Models Understand and Generate Materials
- 2026Zatom-2: Multitask Pretraining on Atomistic Data for Generative Modeling across Domains
- 2026Event-Conditioned Diagnostics of Kinematic, Contact, and Object-Permanence Structure in Passive Object-State World Models
- 2025SwarmThinkers: Learning Physically Consistent Atomic KMC Transitions at Scale
- 2026Baba in Wonderland: Online Self-Supervised Dynamics Discovery for Executable World Models
- 2026PH-Dreamer: A Physics-Driven World Model via Port-Hamiltonian Generative Dynamics