Exploring the robust extrapolation of high-dimensional machine learning potentials

Abstract

We show that, contrary to popular assumptions, predictions from machine learning potentials built upon high-dimensional atom-density representations almost exclusively occur in regions of the representation space which lie outside the convex hull defined by the training set points. We then propose a perspective to rationalize the domain of robust extrapolation and accurate prediction of atomistic machine learning potentials in terms of the probability density induced by training points in the representation space

arXiv comment4 pages, 3 figures

References in corpus (28)

• Gaussian Approximation Potentials: the accuracy of quantum mechanics, without the electrons
• Crystal Graph Convolutional Neural Networks for an Accurate and Interpretable Prediction of Material Properties
• Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning
• SchNet - a deep learning architecture for molecules and materials
• Moment Tensor Potentials: a class of systematically improvable interatomic potentials
• A Spectral Analysis Method for Automated Generation of Quantum-Accurate Interatomic Potentials
• A Performance and Cost Assessment of Machine Learning Interatomic Potentials
• Combining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems
• DScribe: Library of Descriptors for Machine Learning in Materials Science
• Machine Learning Unifies the Modelling of Materials and Molecules
• Active learning of linearly parametrized interatomic potentials
• Physics-inspired structural representations for molecules and materials
• Machine learning a general purpose interatomic potential for silicon
• Estimating the intrinsic dimension of datasets by a minimal neighborhood information
• Ab initio thermodynamics of liquid and solid water
• sGDML: Constructing Accurate and Data Efficient Molecular Force Fields Using Machine Learning
• De novo exploration and self-guided learning of potential-energy surfaces
• Atom-Density Representations for Machine Learning
• Comparison of permutationally invariant polynomials, neural networks, and Gaussian approximation potentials in representing water interactions through many-body expansions
• Uncertainty estimation for molecular dynamics and sampling
• Data-driven simulation and characterisation of gold nanoparticle melting
• A Bayesian approach to NMR crystal structure determination
• Fast Neural Network Approach for Direct Covariant Forces Prediction in Complex Multi-Element Extended Systems
• On Machine Learning Force Fields for Metallic Nanoparticles
• DADApy: Distance-based Analysis of DAta-manifolds in Python
• Generalized convex hull construction for materials discovery
• Simulating solvation and acidity in complex mixtures with first-principles accuracy: the case of CH${}_3$SO${}_3$H and H${}_2$O${}_2$ in phenol
• Compact atomic descriptors enable accurate predictions via linear models

Cited by in corpus (9)

• Systematic Atomic Structure Datasets for Machine Learning Potentials: Application to Defects in Magnesium
• DADApy: Distance-based Analysis of DAta-manifolds in Python
• TimeSOAP: Tracking high-dimensional fluctuations in complex molecular systems via time-variations of SOAP spectra
• "Layer-by-layer" Unsupervised Clustering of Statistically Relevant Fluctuations in Noisy Time-series Data of Complex Dynamical Systems
• Automatic feature selection and weighting in molecular systems using Differentiable Information Imbalance
• Improved decision making with similarity based machine learning: Applications in chemistry
• Uncertainty Quantification in Atomistic Simulations of Silicon using Interatomic Potentials
• Compressing and forecasting atomic material simulations with descriptors
• Assessing zero-shot generalisation behaviour in graph-neural-network interatomic potentials