CrystalJev: thinking fast and slow with atomistic foundation models for materials discovery

ML INF AI4S IMMPP
230
2026年10月04日
原子级基础模型虽能筛选数以百万计的假想材料，却常被当作运行缓慢的模拟器来使用，其经阈值筛选得到的能量值也被直接采信。然而，它们更适合作为快速决策工具来使用。CrystalJev 每次仅对一个未经结构弛豫的晶体结构调用一次冻结的原子间势函数，即可回答各类结构化问题，并输出经校准的概率、有限样本下的理论保证，以及一条明确的“何时应转为慢速思考”的判定规则。在涵盖 65 个 Matbench Discovery 模型的广泛评估中，“稳定”这一二元判断本质上实为一种隐含概率，其具体取值由模型自身的预测误差及候选材料总体的统计分布共同决定。模型一旦训练完成，单次前向传播即可实现接近结构弛豫精度的判别效果，而计算成本仅为后者的约三十分之一；同时，基于信息价值（value-of-information）的理论框架可智能引导：仅当慢速计算确实可能改变最终决策时，才启动该步骤。同一模型层即可统一回答电子结构、力学性能与分子性质等多类问题。在一项已预先注册的前瞻性验证实验中，研究者针对 700 个全新体系开展了密度泛函理论（DFT）计算；结果表明，仅利用既有数据完成校准的单次前向传播预测，在估计未见候选材料中稳定相占比（真实值为 5.8%）时，其高估幅度最大不超过 2.1 个百分点。
Atomistic foundation models triage millions of hypothetical materials but are used as slow simulators, their thresholded energies taken at face value. They are better read as fast decision-makers. CrystalJev queries a frozen interatomic potential once per unrelaxed structure and answers typed questions with calibrated probabilities, finite-sample guarantees and a rule for when to think slowly. Across 65 Matbench Discovery models, a 'stable' call is a probability in disguise, explained by a model's errors and the candidate population. Once trained, one forward pass decides nearly as well as a relaxation at a thirtieth of its cost, and a value-of-information theory sends slower computation only where decisions can change. The same layer answers electronic, mechanical and molecular questions. In a registered prospective test with 700 new density-functional calculations, single-pass forecasts calibrated only on existing data over-stated the stable fraction of unseen candidates (5.8%) by at most 2.1 percentage points.
许愿