Predicting Density of States via Multi-modal Transformer

TL;DR

This paper introduces DOSTransformer, a multi-modal transformer model that predicts the density of states in materials by integrating crystal structure and energy information, outperforming previous methods.

Contribution

The paper presents a novel multi-modal transformer approach for DOS prediction that captures complex relationships between structure and energy levels.

Findings

DOSTransformer outperforms existing models on Phonon and Electron DOS tasks.

The model effectively captures the distribution of states across energy levels.

Extensive experiments validate the model's superiority in real-world scenarios.

Abstract

The density of states (DOS) is a spectral property of materials, which provides fundamental insights on various characteristics of materials. In this paper, we propose a model to predict the DOS by reflecting the nature of DOS: DOS determines the general distribution of states as a function of energy. Specifically, we integrate the heterogeneous information obtained from the crystal structure and the energies via multi-modal transformer, thereby modeling the complex relationships between the atoms in the crystal structure, and various energy levels. Extensive experiments on two types of DOS, i.e., Phonon DOS and Electron DOS, with various real-world scenarios demonstrate the superiority of DOSTransformer. The source code for DOSTransformer is available at https://github.com/HeewoongNoh/DOSTransformer.