Search papers, labs, and topics across Lattice.
CrystalJev reframes atomistic foundation models from slow simulators to fast decision-makers by querying a frozen interatomic potential once per unrelaxed structure to provide calibrated probabilities for material stability. This approach, grounded in value-of-information theory, intelligently directs slower, more expensive computations only when necessary, achieving nearly equivalent decision quality to full relaxations at a fraction of the computational cost. Across 65 Matbench Discovery models and a prospective test, CrystalJev's single-pass forecasts demonstrate accurate, albeit slightly overconfident, predictions of material stability.
A single forward pass through an atomistic foundation model can predict material stability with 30x less computation than a full relaxation, while providing calibrated probabilities and a clear rule for when to "think slowly."
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.
