Direct prediction of phonon density of states with Euclidean neural networks

Abstract

Machine learning has demonstrated great power in materials design, discovery, and property prediction. However, despite the success of machine learning in predicting discrete properties, challenges remain for continuous property prediction. The challenge is aggravated in crystalline solids due to crystallographic symmetry considerations and data scarcity. Here we demonstrate the direct prediction of phonon density of states using only atomic species and positions as input. We apply Euclidean neural networks, which by construction are equivariant to 3D rotations, translations, and inversion and thereby capture full crystal symmetry, and achieve high-quality prediction using a small training set of $\sim 10^3$ examples with over 64 atom types. Our predictive model reproduces key features of experimental data and even generalizes to materials with unseen elements,and is naturally suited to efficiently predict alloy systems without additional computational cost. We demonstrate the potential of our network by predicting a broad number of high phononic specific heat capacity materials. Our work indicates an efficient approach to explore materials' phonon structure, and can further enable rapid screening for high-performance thermal storage materials and phonon-mediated superconductors.

21 pages total, 5 main figures + 16 supplementary figures. To appear in Advanced Science (2021)

References in corpus (2)

• A General-Purpose Machine Learning Framework for Predicting Properties of Inorganic Materials
• Origin of the large phonon band-gap in SrTiO3 and the vibrational signatures of ferroelectricity in ATiO3 perovskite: First principles lattice dynamics and inelastic neutron scattering of PbTiO3, BaTiO3 and SrTiO3

Cited by in corpus (14)

• Recent Advances and Applications of Deep Learning Methods in Materials Science
• A Physics-Guided Neural Network Framework for Elastic Plates: Comparison of Governing Equations-Based and Energy-Based Approaches
• Machine Learning on Neutron and X-Ray Scattering
• Theory for Equivariant Quantum Neural Networks
• Density of States Prediction for Materials Discovery via Contrastive Learning from Probabilistic Embeddings
• Formula graph self-attention network for representation-domain independent materials discovery
• Robust Machine Learning Inference from X-ray Absorption Near Edge Spectra through Featurization
• Surrogate- and invariance-boosted contrastive learning for data-scarce applications in science
• Representations and Strategies for Transferable Machine Learning Models in Chemical Discovery
• Artificial Intelligence for Science in Quantum, Atomistic, and Continuum Systems
• Decoding Structure-Spectrum Relationships with Physically Organized Latent Spaces
• Mechanism of Local Lattice Distortion Effects on Vacancy Migration Barriers in FCC Alloys
• Accelerating superconductor discovery through tempered deep learning of the electron-phonon spectral function
• Crystalline Material Discovery in the Era of Artificial Intelligence