AI RESEARCH
Proof-Carrying Materials: Falsifiable Safety Certificates for Machine-Learned Interatomic Potentials
arXiv CS.AI
•
ArXi:2603.12183v1 Announce Type: cross Machine-learned interatomic potentials (MLIPs) are deployed for high-throughput materials screening without formal reliability guarantees. We show that a single MLIP used as a stability filter misses 93% of density functional theory (DFT)-stable materials (recall 0.07) on a 25,000-material benchmark. Proof-Carrying Materials (PCM) closes this gap through three stages: adversarial falsification across compositional space, bootstrap envelope refinement with 95% confidence intervals, and Lean 4 formal certification.