Abstract:Deep learning neural language model is used to learn the relationships of materials compositions and elemental oxidation states and exploit this deep dark knowledge for accurate prediction of atomic site oxidation states for materials given only their compositions. This model can be used to speed up high‐throughput generative discovery of new materials. Oxidation states (OS) are the charges on atoms due to electrons gained or lost upon applying an ionic approximation to their bonds. As a fundamental property, OS has been widely used in charge‐neutrality verification, crystal structure determination, and reaction estimation. Currently, only heuristic rules exist for guessing the oxidation states of a given compound with many exceptions. Recent work has developed machine learning models based on heuristic structural features for predicting the oxidation states of metal ions. However, composition‐based oxidation state prediction still remains elusive so far, which has significant implications for the discovery of new materials for which the structures have not been determined. This work proposes a novel deep learning‐based BERT transformer language model BERTOS for predicting the oxidation states for all elements of inorganic compounds given only their chemical composition. This model achieves 96.82% accuracy for all‐element oxidation states prediction benchmarked on the cleaned ICSD dataset and achieves 97.61% accuracy for oxide materials. It is also demonstrated how it can be used to conduct large‐scale screening of hypothetical material compositions for materials discovery.

Ionic species representations for materials informatics

Element similarity in high-dimensional materials representations

Spectral Operator Representations

Mapping inorganic crystal chemical space

Data-Mining Element Charges in Inorganic Materials

Persona of Transition Metal Ions in Solids: A Statistical Learning on Local Structures of Transition Metal Oxides

High-throughput calculations and machine learning modeling of 17O NMR in non-magnetic oxides

Exploring material compositions for synthesis using oxidation states

Distributed Representations of Atoms and Materials for Machine Learning

Material Property Prediction with Element Attribute Knowledge Graphs and Multimodal Representation Learning

Structure to Property: Chemical Element Embeddings for Predicting Electronic Properties of Crystals

Discovering High-Entropy Oxides with a Machine-Learning Interatomic Potential

Structure motif centric learning framework for inorganic crystalline systems

Composition based oxidation state prediction of materials using deep learning

Composition Based Oxidation State Prediction of Materials Using Deep Learning Language Models

Leveraging language representation for materials exploration and discovery

Data-driven Discovery of Photoactive Quaternary Oxides using First-principles Machine Learning

From structure mining to unsupervised exploration of atomic octahedral networks

Extending Shannon's Ionic Radii Database Using Machine Learning

IonML: a physically inspired machine learning platform to directed design superionic conductors