Materials property prediction using symmetry-labeled graphs as atomic-position independent descriptors
arXiv:1905.06048 · doi:10.1103/PhysRevB.100.104114
Abstract
Computational materials screening studies require fast calculation of the properties of thousands of materials. The calculations are often performed with Density Functional Theory (DFT), but the necessary computer time sets limitations for the investigated material space. Therefore, the development of machine learning models for prediction of DFT calculated properties are currently of interest. A particular challenge for \emph{new} materials is that the atomic positions are generally not known. We present a machine learning model for the prediction of DFT-calculated formation energies based on Voronoi quotient graphs and local symmetry classification without the need for detailed information about atomic positions. The model is implemented as a message passing neural network and tested on the Open Quantum Materials Database (OQMD) and the Materials Project database. The test mean absolute error is 20 meV on the OQMD database and 40 meV on Materials Project Database. The possibilities for prediction in a realistic computational screening setting is investigated on a dataset of 5976 ABSe selenides with very limited overlap with the OQMD training set. Pretraining on OQMD and subsequent training on 100 selenides result in a mean absolute error below 0.1 eV for the formation energy of the selenides.
arXiv comment14 pages including references and 13 figures
References in corpus (16)
- Restoring the density-gradient expansion for exchange in solids and surfaces
- Generalized gradient approximation for solids and their surfaces
- Deep Residual Learning for Image Recognition
- Crystal Graph Convolutional Neural Networks for an Accurate and Interpretable Prediction of Material Properties
- SchNet - a deep learning architecture for molecules and materials
- Two-dimensional materials from high-throughput computational exfoliation of experimentally known compounds
- Quantum-Chemical Insights from Deep Tensor Neural Networks
- Universal Fragment Descriptors for Predicting Electronic Properties of Inorganic Crystals
- How to represent crystal structures for machine learning: towards fast prediction of electronic properties
- Identification of structure in condensed matter with the topological cluster classification
- A Topological Framework for Local Structure Analysis in Condensed Matter
- Randomized Prior Functions for Deep Reinforcement Learning
- Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials
- Atomic positions independent descriptor for machine learning of material properties
- Generating equilibrium molecules with deep neural networks
- Bayesian Inference with Anchored Ensembles of Neural Networks, and Application to Exploration in Reinforcement Learning