DeepH-2: Enhancing deep-learning electronic structure via an equivariant local-coordinate transformer

2024/01/30 by Yuxiang Wang, Wang, Yuxiang, He Li +15 · 3 citations
Computer Science · #Parallel Computing and Optimization Techniques

paper · pdf · doi:10.48550/arxiv.2401.17015

Abstract

Deep-learning electronic structure calculations show great potential for revolutionizing the landscape of computational materials research. However, current neural-network architectures are not deemed suitable for widespread general-purpose application. Here we introduce a framework of equivariant local-coordinate transformer, designed to enhance the deep-learning density functional theory Hamiltonian referred to as DeepH-2. Unlike previous models such as DeepH and DeepH-E3, DeepH-2 seamlessly integrates the simplicity of local-coordinate transformations and the mathematical elegance of equivariant neural networks, effectively overcoming their respective disadvantages. Based on our comprehensive experiments, DeepH-2 demonstrates superiority over its predecessors in both efficiency and accuracy, showcasing state-of-the-art performance. This advancement opens up opportunities for exploring universal neural network models or even large materials models.

Cited by

• Predicting the Thermal Behavior of Semiconductor Defects with Equivariant Neural Networks
• Advancing Universal Deep Learning for Electronic-Structure Hamiltonian Prediction of Materials
• Hot-Ham: an accurate and efficient E(3)-equivariant machine-learning electronic structures calculation framework

Related

• Deep learning density functional theory Hamiltonian in real space
• Efficient hybrid density functional calculation by deep learning
• Hot-Ham: an accurate and efficient E(3)-equivariant machine-learning electronic structures calculation framework
• Bridging deep learning force fields and electronic structures with a physics-informed approach
• Advancing Universal Deep Learning for Electronic-Structure Hamiltonian Prediction of Materials