Density of States Prediction for Materials Discovery via Contrastive Learning from Probabilistic Embeddings

Abstract

Machine learning for materials discovery has largely focused on predicting an individual scalar rather than multiple related properties, where spectral properties are an important example. Fundamental spectral properties include the phonon density of states (phDOS) and the electronic density of states (eDOS), which individually or collectively are the origins of a breadth of materials observables and functions. Building upon the success of graph attention networks for encoding crystalline materials, we introduce a probabilistic embedding generator specifically tailored to the prediction of spectral properties. Coupled with supervised contrastive learning, our materials-to-spectrum (Mat2Spec) model outperforms state-of-the-art methods for predicting ab initio phDOS and eDOS for crystalline materials. We demonstrate Mat2Spec's ability to identify eDOS gaps below the Fermi energy, validating predictions with ab initio calculations and thereby discovering candidate thermoelectrics and transparent conductors. Mat2Spec is an exemplar framework for predicting spectral properties of materials via strategically incorporated machine learning techniques.

26 pages, 15 figures, Minor edits to match the proofs from Nat.Commun

References in corpus (3)

• A General-Purpose Machine Learning Framework for Predicting Properties of Inorganic Materials
• Disentangled Variational Autoencoder based Multi-Label Classification with Covariance-Aware Multivariate Probit Model
• Deep Hurdle Networks for Zero-Inflated Multi-Target Regression: Application to Multiple Species Abundance Estimation

Cited by in corpus (6)

• A critical examination of robustness and generalizability of machine learning prediction of materials properties
• Multimodal Foundation Models for Material Property Prediction and Discovery
• Machine Learning and Data-Driven Methods in Computational Surface and Interface Science
• Adaptive energy reference for machine-learning models of the electronic density of states
• Self-Supervised Generative Models for Crystal Structures
• Predicting Many Crystal Properties via an Adaptive Transformer-based Framework