Proof-Carrying Materials: Falsifiable Safety Certificates for Machine-Learned Interatomic Potentials
Not on the map. Why?
Abstract (arXiv)
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. Auditing CHGNet, TensorNet and MACE reveals architecture-specific blind spots with near-zero pairwise error correlations (r <= 0.13; n = 5,000), confirmed by independent Quantum ESPRESSO validation (20/20 converged; median DFT/CHGNet force ratio 12x). A risk model trained on PCM-discovered features predicts failures on unseen materials (AUC-ROC = 0.938 +/- 0.004) and transfers across architectures (cross-MLIP AUC-ROC ~ 0.70; feature importance r = 0.877). In a thermoelectric screening case study, PCM-audited protocols discover 62 additional stable materials missed by single-MLIP screening - a 25% improvement in discovery yield.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 24 of its 68 references that are in the map. First posted on arXiv 12 Mar 2026.
Builds on
24 of 68 references.
References · 24
- 2014Explaining and Harnessing Adversarial Examplescited by 7,364
- 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
- 2015Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learningcited by 4,448
- 2016Simple and Scalable Predictive Uncertainty Estimation using Deep Ensemblescited by 3,926
- 2016Concrete Problems in AI Safetycited by 1,642
- 2017Reluplex: An Efficient SMT Solver for Verifying Deep Neural Networkscited by 744
- 2022MACE: Higher Order Equivariant Message Passing Neural Networks for Fast and Accurate Force Fieldscited by 474
- 2022Conformal Risk Controlcited by 299
- 2025DeepSeek-Prover-V2: Advancing Formal Mathematical Reasoning via Reinforcement Learning for Subgoal Decompositioncited by 287
- 2020Fast and Uncertainty-Aware Directional Message Passing for Non-Equilibrium Moleculescited by 255
- 2023EquiformerV2: Improved Equivariant Transformer for Scaling to Higher-Degree Representationscited by 204
- 2024Towards Guaranteed Safe AI: A Framework for Ensuring Robust and Reliable AI Systemscited by 96
12 more
- 202314 Examples of How LLMs Can Transform Materials Science and Chemistry: A Reflection on a Large Language Model Hackathoncited by 86
- 2018Less is more: sampling chemical space with active learningcited by 82
- 2020Uncertainty Quantification Using Neural Networks for Molecular Property Predictioncited by 66
- 2021Guidance on the Assurance of Machine Learning in Autonomous Systems (AMLAS)cited by 42
- 2024Are LLMs Ready for Real-World Materials Discovery?cited by 40
- 2023Matbench Discovery -- A framework to evaluate machine learning crystal stability predictionscited by 35
- 2017Compositional Falsification of Cyber-Physical Systems with Machine Learning Componentscited by 32
- 2023LLM-Prop: Predicting Physical And Electronic Properties Of Crystalline Solids From Their Text Descriptionscited by 29
- 2025A practical guide to machine learning interatomic potentials -- Status and futurecited by 26
- 2020Committee neural network potentials control generalization errors and enable active learningcited by 15
- 2025Learning atomic forces from uncertainty-calibrated adversarial attackscited by 2
- 2025Proof-Carrying Neuro-Symbolic Codecited by 1
Built on by
Citing papers per year: 2026: 1.
1 paper.