Putting Density Functional Theory to the Test in Machine-Learning-Accelerated Materials Discovery

Abstract

Accelerated discovery with machine learning (ML) has begun to provide the advances in efficiency needed to overcome the combinatorial challenge of computational materials design. Nevertheless, ML-accelerated discovery both inherits the biases of training data derived from density functional theory (DFT) and leads to many attempted calculations that are doomed to fail. Many compelling functional materials and catalytic processes involve strained chemical bonds, open-shell radicals and diradicals, or metal-organic bonds to open-shell transition-metal centers. Although promising targets, these materials present unique challenges for electronic structure methods and combinatorial challenges for their discovery. In this Perspective, we describe the advances needed in accuracy, efficiency, and approach beyond what is typical in conventional DFT-based ML workflows. These challenges have begun to be addressed through ML models trained to predict the results of multiple methods or the differences between them, enabling quantitative sensitivity analysis. For DFT to be trusted for a given data point in a high-throughput screen, it must pass a series of tests. ML models that predict the likelihood of calculation success and detect the presence of strong correlation will enable rapid diagnoses and adaptation strategies. These "decision engines" represent the first steps toward autonomous workflows that avoid the need for expert determination of the robustness of DFT-based materials discoveries.

References in corpus (21)

• UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction
• Big Data meets Quantum Chemistry Approximations: The $\Delta$-Machine Learning Approach
• By-passing the Kohn-Sham equations with machine learning
• Machine learning prediction errors better than DFT accuracy
• Automated Selection of Active Orbital Spaces
• Resolving transition metal chemical space: feature selection for machine learning and structure-property relationships
• Fractional spins and static correlation error in density functional theory
• Transferability in Machine Learning for Electronic Structure via the Molecular Orbital Basis
• Predicting Electronic Structure Properties of Transition Metal Complexes with Neural Networks
• Deep learning and the Schrödinger equation
• Fast and Accurate Uncertainty Estimation in Chemical Machine Learning
• Can single-reference coupled cluster theory describe static correlation?
• Understanding the mechanism stabilizing intermediate spin states in Fe(II)-Porphyrin
• Assessment of density functional theory for iron(II) molecules across the spin-crossover transition
• Separation of Dynamic and Nondynamic Correlation
• Towards quantifying the role of exact exchange in predictions of transition metal complex properties
• Parallel and Distributed Thompson Sampling for Large-scale Accelerated Exploration of Chemical Space
• Semiempirical Molecular Orbital Models based on the Neglect of Diatomic Differential Overlap Approximation
• Local descriptors of dynamic and nondynamic correlation
• The Shape of Full Configuration Interaction to Come
• Do CCSD and approximate CCSD-F12 variants converge to the same basis set limits? The case of atomization energies

Cited by in corpus (4)

• Performance of the r${}^2$SCAN functional in transition metal oxides
• The long road to calibrated prediction uncertainty in computational chemistry
• Molecular-orbital-based Machine Learning for Open-shell and Multi-reference Systems with Kernel Addition Gaussian Process Regression
• Electronic structures of crystalline and amorphous GeSe and GeSbTe compounds using machine learning empirical pseudopotentials